1: %#! latex irreg2
2: \documentclass{amsart}
3: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
4: \newcommand{\timestamp}{\today}
5: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
6: \usepackage{amsthm}
7: %%\usepackage{showkeys}
8: \usepackage{epic,eepic}
9: \usepackage[dvips]{graphicx}
10: \renewcommand{\subjclassname}{\textup{2000} Mathematics Subject
11: Classification}
12: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
13: \makeatletter
14: \def\authornote#1{%
15: \marginpar{\footnotesize #1}%
16: %% \else\ignorespaces%
17: \relax%
18: }
19: \newcounter{tempcounteri}
20: \newcounter{tempcounterii}
21: \newenvironment{enumeraterom}{%
22: \begin{enumerate}%
23: \renewcommand{\theenumi}{(\roman{enumi})}
24: \renewcommand{\labelenumi}{(\roman{enumi})}}{%
25: \end{enumerate}}
26: \renewcommand{\theenumi}{(\arabic{enumi})}
27: \renewcommand{\labelenumi}{\theenumi}
28: \newcommand{\bmath}[1]{\mbox{\mathversion{bold}$#1$}}
29: \newcommand{\smath}[1]{\mbox{\mathversion{bold}\scriptsize{$#1$}}}
30: \newcommand{\C}{\bmath{C}}
31: \newcommand{\subC}{\smath{C}}
32: \newcommand{\Z}{\bmath{Z}}
33: \newcommand{\R}{\bmath{R}}
34: \newcommand{\subR}{\smath{R}}
35: \newcommand{\CP}{\bmath{C\!P}}
36: \newcommand{\SL}{\operatorname{SL}}
37: \newcommand{\SU}{\operatorname{SU}}
38: \newcommand{\U}{\operatorname{U}}
39: \newcommand{\PSL}{\operatorname{PSL}}
40: \newcommand{\PSU}{\operatorname{PSU}}
41: \newcommand{\Herm}{\operatorname{Herm}}
42: \newcommand{\Trig}{\operatorname{{\mathcal T}}}
43: \newcommand{\cmcone}{\mbox{\rm CMC}-$1$}
44: \newcommand{\TA}{\operatorname{TA}}
45: \newcommand{\trace}{\operatorname{trace}}
46: \newcommand{\ord}{\operatorname{ord}}
47: \newcommand{\id}{\operatorname{id}}
48: \newcommand{\Hyp}{\mathcal{H}}
49: \newcommand{\metone}{\operatorname{Met}_1}
50: \renewcommand{\Re}{\operatorname{Re}}
51: \renewcommand{\Im}{\operatorname{Im}}
52: \newcommand{\Res}{\operatornamewithlimits{Res}}
53: %%% Theorems
54: \newtheorem{theorem}{Theorem}[section]
55: \newtheorem*{theorem*}{Theorem}
56: \newtheorem{proposition}[theorem]{Proposition}
57: \newtheorem{corollary}[theorem]{Corollary}
58: \newtheorem{lemma}[theorem]{Lemma}
59: \theoremstyle{definition}
60: \newtheorem{definition}[theorem]{Definition}
61: \newtheorem{example}[theorem]{Example}
62: \theoremstyle{remark}
63: \newtheorem{remark}[theorem]{Remark}
64: \newtheorem*{remark*}{Remark}
65: \numberwithin{equation}{section}
66: \setcounter{tocdepth}{1}
67: \makeatother
68: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
69: %% Top Matters
70: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
71: \title[Symmetric CMC-1 surfaces with irregular ends]{
72: Constructing mean curvature $1$ surfaces \\
73: in $H^3$ with irregular ends
74: }
75:
76: \author{Wayne Rossman}
77: \author{Masaaki Umehara}
78: \author{Kotaro Yamada}
79: \date{December 15, 2002}
80: \address[Rossman]{%
81: Department of Mathematics, Faculty of Science,
82: Kobe University,
83: Rokko, Kobe 657-8501, Japan%
84: }
85: \email{wayne@math.kobe-u.ac.jp}
86: \address[Umehara]{%
87: Department of Mathematics, Graduate School of Science,
88: Osaka University,
89: Osaka 560-0043, Japan%
90: }
91: \email{umehara@math.sci.osaka-u.ac.jp}
92: \address[Yamada]{%
93: Faculty of Mathematics,
94: Kyushu University 36, 6-10-1
95: Hakozaki, Higashi-ku, Fukuoka 812-8581, Japan%
96: }
97: \email{kotaro@math.kyushu-u.ac.jp}
98: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
99: \begin{document}
100: \begin{abstract}
101: With the developments of the last decade on complete constant mean
102: curvature $1$ (\cmcone) surfaces in the hyperbolic $3$-space $H^3$, many
103: examples of such surfaces are now known.
104: However, most of the known examples have regular ends.
105: (An end is irregular, resp.~regular, if the hyperbolic Gauss map of the
106: surface has an essential singularity, resp.~at most a pole, there.)
107: There are some known surfaces with irregular ends,
108: but they are all either reducible or of infinite total curvature.
109: (The surface is reducible if and only if the monodromy of the
110: secondary Gauss map can be simultaneously diagonalized.)
111: Up to now there have been no known complete irreducible \cmcone{} surfaces
112: in $H^3$ with finite total curvature and irregular ends.
113:
114: The purpose of this paper is to construct countably many $1$-parameter
115: families of genus zero \cmcone{} surfaces with irregular ends and finite
116: total curvature, which have either dihedral or Platonic symmetries.
117: For all the examples we produce, we show that they have finite total
118: curvature and irregular ends.
119: For the examples with dihedral symmetry and the simplest
120: example with tetrahedral symmetry, we show irreducibility.
121: Moreover, we construct a genus one \cmcone{} surface with four irregular
122: ends, which is the first known example with positive genus whose ends
123: are all irregular.
124: \end{abstract}
125: \maketitle
126: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
127: \section*{Introduction}
128: Let $H^3$ denote the unique simply connected complete $3$-dimensional
129: Riemannian manifold with constant sectional curvature $-1$, which we
130: call the hyperbolic $3$-space.
131: Associated to a complete finite-total-curvature \cmcone{}
132: (constant mean curvature one)
133: conformal immersion $f \colon{}M \to H^3$
134: of a Riemann surface $M$ are two meromorphic maps called the hyperbolic
135: Gauss map and the secondary Gauss map, which we denote by $G$ and $g$
136: respectively (to be defined in the next section).
137: Using these two Gauss maps, we can define two characteristics of the
138: surface $f$:
139: \begin{enumerate}
140: \item It is known that $M$ is biholomorphic to a compact Riemann surface
141: with a finite number of points removed, and hence each end is
142: conformally a punctured disk.
143: Therefore we may consider the order of the hyperbolic Gauss map
144: $G$ at each end, and an end is called {\it regular\/} if $G$ has
145: at most a pole singularity at this end.
146: If $G$ has an essential singularity, the end is called
147: {\it irregular}.
148: \item Although $G$ is single-valued on $M$, the secondary Gauss map
149: $g$ might be multi-valued on $M$, so we can have a nontrivial
150: monodromy representation defined on the first fundamental group of
151: $M$.
152: This monodromy group is a subgroup of $\SU(2)$, and if all members
153: of this group can be diagonalized by the same conjugation, we say
154: that the surface $f$ is {\it reducible}.
155: Otherwise, we say that $f$ is {\it irreducible}.
156: (Irreducibility depends on a global behavior of the surface but
157: not on individual ends.)
158: \end{enumerate}
159: If a \cmcone{} immersion is reducible, the surface can be deformed
160: preserving its hyperbolic Gauss map $G$ and Hopf differential
161: ($(2,0)$-part of the second fundamental form, see
162: Section~\ref{sec:prelim}).
163: On the other hand, an irreducible surface is the only surface with given
164: hyperbolic Gauss map and Hopf differential.
165:
166: Recent progress in the theory of \cmcone{} surfaces in $H^3$ has led to
167: the discovery of many new examples of these surfaces.
168: Many examples with regular ends are now known, and various properties of
169: these surfaces are understood.
170: Bryant \cite{Bryant} found a local description for these surfaces in terms
171: of holomorphic data that initiated this recent progress.
172: The last two authors \cite{uy1}--\cite{uy7} developed the theory using
173: Bryant's description to find many examples and properties, and work in
174: this direction has been continued by Small \cite{Sm}, the authors
175: \cite{ruy1}--\cite{ruy5}, Costa-Sousa Neto \cite{CN},
176: Earp-Toubiana \cite{ET1}--\cite{ET2}, Yu \cite{Yu1}--\cite{Yu3},
177: Levi-Rossman \cite{LR}, Barbosa-Berard \cite{BB}, do Carmo and Gomes and
178: Lawson and Thorbergsson and Silveira \cite{CGT,CL,CS}, and others.
179: Regarding properties of the ends of embedded examples, Collin, Hauswirth
180: and Rosenberg \cite{chr} have recently shown that any embedded \cmcone{}
181: surface of finite total curvature is either a horosphere or all of its
182: ends are asymptotic to catenoid cousin ends.
183: In \cite{chr,Yu3} it is further shown that any irregular end cannot
184: be embedded, and the limit points of such an end are dense at infinity.
185: Recently, Pacard and Pimentel \cite{PP} established a method for
186: attaching small handles between tangent horospheres and deforming to
187: produce \cmcone{} surfaces, and this construction produces many embedded
188: \cmcone{} surfaces of any genus.
189: Also, Karcher \cite{Kar} has recently constructed periodic \cmcone{}
190: surfaces with fundamental domains in several different types of compact
191: quotients of $H^3$.
192:
193:
194: A typical example of an irregular end is the end of the Enneper cousin,
195: a surface first constructed by Bryant \cite{Bryant}.
196: After that the last two authors \cite{uy2} constructed examples of genus
197: zero and two irregular ends, and also many reducible \cmcone{} surfaces
198: of genus zero whose ends are all irregular, using deformations from
199: minimal surfaces.
200: (The conclusion of Remark 4.4 in \cite{uy2} contains an error.
201: The number of ends should be $ml+2$, and hence the genus of
202: $\hat M_0^*$ is zero.) Recently, Daniel \cite{D} has investigated
203: irregular ends
204: from the viewpoint of Nevanlinna theory.
205:
206: After \cite{uy2}, no further surfaces with irregular ends and
207: finite total curvature had been constructed.
208: (However, such an example with infinite total curvature can be found in
209: \cite{ruy3}.)
210: In particular, until now no irreducible \cmcone{} surfaces with
211: irregular ends and finite total curvature had been known.
212:
213: The purpose of the paper is to construct countably many $1$-parameter
214: families of genus zero \cmcone{} surfaces with irregular ends and finite
215: total curvature, which have either dihedral or Platonic symmetries.
216: We further show that examples with dihedral symmetries, and the simplest
217: example with tetrahedral symmetry, must be irreducible.
218: All of our examples have
219: irregular ends of finite type in the sense of Daniel \cite{D}.
220:
221: To do the construction, we start with the meromorphic data for the genus
222: zero irreducible \cmcone{} surfaces with regular ends found in
223: \cite{uy3} and \cite{ruy1} and modify this data to make surfaces with
224: irregular ends.
225: The spirit of the construction is similar to the construction of
226: trinoids in \cite{uy3}, where \cmcone{} surfaces with prescribed Gauss
227: maps are constructed by reflecting spherical triangles, and we use
228: monodromy killing arguments like in \cite{ruy1} and \cite{uy6}, but
229: the techniques are brought to bear more intricately here.
230:
231: In Section~\ref{sec:prelim} we give necessary preliminaries.
232: As our construction is done by reflecting abstract spherical triangles,
233: we discuss this in Section~\ref{sec:reflection}, and introduce a method
234: to construct \cmcone{} surfaces with irregular ends
235: (Theorem~\ref{thm:gen}), which is proved in
236: Section~\ref{sec:proof}.
237: As an application of the theorem, we construct examples of genus zero
238: with either dihedral or Platonic symmetries in Section~\ref{sec:platonic}.
239: Finally, in Section~\ref{sec:torus}, we construct a \cmcone{} surface of
240: genus $1$ with four irregular ends, which is the first known example
241: with positive genus whose ends are all irregular.
242: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
243: \section{Preliminaries}
244: \label{sec:prelim}
245: \subsection*{Null meromorphic curves.}
246: Here we recall from \cite{uy1,uy7} some fundamental properties of null
247: meromorphic curves in $\SL(2,\C)$.
248: \begin{definition}
249: Let $F\colon{}M \to \SL(2,\C)$ be a meromorphic map defined on a
250: Riemann surface $M$ with a local complex coordinate $z$.
251: Then $F$ is called {\it null\/} if $\det(F_z) \equiv 0$.
252: (This condition does not depend on the choice of coordinate $z$.)
253: \end{definition}
254: Let $F\colon{}M \to \SL(2,\C)$ be a null meromorphic map.
255: We define a matrix $\alpha$ by
256: \[
257: \alpha=\begin{pmatrix} \alpha_{11} & \alpha_{12} \\
258: \alpha_{21} & \alpha_{22}
259: \end{pmatrix}
260: :=F^{-1}dF \; ,
261: \]
262: and set
263: \begin{equation}\label{eq:gomega}
264: g:= \alpha_{11}/\alpha_{21}\;,\qquad
265: \omega:=\alpha_{21} \; .
266: \end{equation}
267: Then the pair $(g,\omega)$ is a meromorphic function $g$ and a
268: meromorphic $1$-form $\omega$ on $M$ satisfying
269: \begin{equation}\label{eq:ode0}
270: F^{-1}dF= \begin{pmatrix}
271: g & -g^2 \\
272: 1 & -g\hphantom{^2}
273: \end{pmatrix} \, \omega \; .
274: \end{equation}
275: Conversely, let $g$ be a meromorphic function and $\omega$ a holomorphic
276: $1$-form on $M$.
277: Then the ordinary differential equation \eqref{eq:ode0} is integrable
278: and the solution $F$ is a null map into $\SL(2,\C)$ (since we will always
279: choose the initial condition to be in $\SL(2,\C)$) defined on the
280: universal covering of $M\setminus\{\text{poles of $\alpha$}\}$.
281: In general, $F$ might not be single-valued on $M$ itself, and $F$ may have
282: essential singularities at poles of $\alpha$.
283: We call the pair $(g,\omega)$ the {\it Weierstrass data\/} of $F$.
284: \begin{definition}
285: Let
286: \[
287: F=\begin{pmatrix}F_{11} & F_{12} \\ F_{21} & F_{22}
288: \end{pmatrix}
289: \]
290: be a null meromorphic map of $M$ into $\SL(2,\C)$ with
291: Weierstrass data $(g,\omega)$.
292: We call
293: \[
294: G:=\frac {dF_{11}}{dF_{21}}=\frac {dF_{12}}{dF_{22}}
295: \]
296: the {\it hyperbolic Gauss map\/} of $F$.
297: Furthermore, we call $g$ in \eqref{eq:gomega}
298: the {\it secondary Gauss map\/} and $Q=\omega\cdot dg$
299: the {\it Hopf differential\/} of $F$.
300: \end{definition}
301: We remark that the secondary Gauss map $g$ satisfies
302: \[
303: g=-\frac {dF_{22}}{dF_{21}}=-\frac {dF_{12}}{dF_{11}} \; .
304: \]
305:
306: Let $F\colon{} M \to \SL(2,\C)$ be a null meromorphic map.
307: Then for $a,b\in \SL(2,\C)$, $\widetilde F =aFb^{-1}$ is also a null
308: meromorphic map.
309: Then the associated two Gauss maps $\widetilde G$, $\widetilde g$, and
310: the Hopf differential $\widetilde Q$ of $\widetilde F$ are
311: \begin{equation}\label{eq:three}
312: \widetilde G=a\star G \; , \quad
313: \widetilde g=b\star g \; , \quad
314: \mbox{and}\quad
315: \widetilde Q=Q \; ,
316: \end{equation}
317: where, for any matrix $a=(a_{ij})\in\SL(2,\C)$ and any function $G$,
318: $a\star G$ is the M\"obius transformation of $G$ by $a$:
319: \begin{equation}\label{eq:moebius}
320: a\star G = \frac{a_{11}G+a_{12}}{a_{21}G+a_{22}} \; .
321: \end{equation}
322:
323: Let $z$ be a complex coordinate on a neighborhood $U$ of $M$.
324: Now we consider the Schwarzian derivatives $S(G)$ and $S(g)$ on $U$ of
325: $G$ and $g$, where
326: \begin{equation}\label{eq:sch}
327: S(G)=
328: \left[
329: \left(\frac{G''}{G'}\right)'-\frac 12 \left(\frac{G''}{G'}\right)^2
330: \right]\,dz^2 \qquad
331: \left( '=\frac{d}{dz}\right).
332: \end{equation}
333: The description of the Schwarzian derivative depends on the choice of
334: complex coordinate $z$.
335: However, any difference of two Schwarzian derivatives, as a meromorphic
336: $2$-differential, is independent of the choice of complex coordinate.
337: Note that the Schwarzian derivative is invariant under M\"obius
338: transformations:
339: \begin{equation}\label{eq:sch-inv}
340: S(G)=S(a\star G) \qquad \bigl(a\in\SL(2,\C)\bigr).
341: \end{equation}
342: The following identity can be checked:
343: \begin{equation}\label{S-rel}
344: S(g)-S(G)=2Q\;.
345: \end{equation}
346: Conversely, the following lemma holds:%
347: \begin{lemma}[\cite{Sm,uy3,kuy}]\label{lem:unique}
348: Let $(G,g)$ be a pair of meromorphic functions on $M$ such that
349: $S(g)-S(G)$ is not identically zero.
350: Then there exists a unique {\rm (}up to sign{\rm )} null meromorphic
351: map $F\colon{}M\to \SL(2,\C)$ such that $G$ and $g$ are the
352: hyperbolic Gauss map and the secondary Gauss map of $F$.
353: \end{lemma}
354:
355: Now, for later use, we point out the following elementary fact from
356: linear algebra:
357: \begin{lemma}[\cite{ruy1}]\label{fact}
358: A matrix $a\in \SL(2,\C)$ satisfies $a\bar a=\id$
359: if and only if it is of the form
360: \[
361: a=\begin{pmatrix}
362: p & i \gamma_1 \\
363: i\gamma_2 & \bar p
364: \end{pmatrix}
365: \qquad \text{with}\quad \gamma_1,\gamma_2\in \R \; , \quad
366: p\bar p+\gamma_1\gamma_2=1 \; , \quad i=\sqrt{-1} \; .
367: \]
368: \end{lemma}
369:
370: \subsection*{CMC-1 surfaces in $H^3$}
371: We identify the Minkowski 4-space $L^4$, which has the canonical
372: Lorentzian metric $(~\cdot~,~\cdot~)$ of signature $(-,+,+,+)$,
373: with the space of $2\times 2$ hermitian matrices $\Herm(2)$.
374: More explicitly, $(t,x_1,x_2,x_3)\in L^4$ is identified with the matrix
375: \[
376: \begin{pmatrix}
377: t+x_3 & x_1+ix_2 \\
378: x_1-ix_2& t-x_3
379: \end{pmatrix}
380: \in \Herm(2) \; .
381: \]
382: The hyperbolic $3$-space can be defined as the upper component
383: \[ H^3=\{\xi=(t,x_1,x_2,x_3) \in L^4\,|\,(\xi,\xi)=-1 \,,\;t > 0\,\} \]
384: of the hyperboloid in $L^4$ with the
385: induced metric.
386: In $\Herm(2)$ this is represented as
387: \[
388: H^3=\{\,X\in \Herm(2)\,;\,\det X=1 \, ,\,\, \trace X >0\,\}
389: =\{aa^*\,;\,a\in\SL(2,\C)\} \;,
390: \]
391: where $a^*={}^t\bar a$.
392: The complex Lie group $\SL(2,\C)$ acts isometrically on $H^3$ by
393: $\rho(a)x=a x a^*$, where $a\in \SL(2,\C)$ and $x\in H^3$.
394:
395: Let $M$ be a Riemann surface and $F\colon{}M\to \SL(2,\C)$ a null
396: holomorphic immersion.
397: Then $f=FF^*\colon{}M\to H^3$ is a conformal \cmcone{} immersion.
398: Conversely, for any conformal \cmcone{} immersion $f\colon{}M\to H^3$,
399: there exists a null holomorphic immersion $F\colon{}\widetilde M\to
400: \SL(2,\C)$ such that $f=FF^*$,
401: where $\widetilde M$ denotes the universal covering of $M$.
402: We call $F$ a {\it lift\/} of the conformal \cmcone{} immersion $f$.
403: Let $\widetilde F$ be another lift of $f$.
404: Then, we have the expression $\widetilde F=Fb^{-1}$ for some
405: matrix $b\in \SU(2)$.
406: Let $(g,\omega)$ be the Weierstrass data of the null map of the lift
407: $F$.
408: Then the first fundamental form $ds^2$ and the second fundamental form
409: $I\!I$ are given by (see \cite{uy1}, for example)
410: \begin{equation}\label{eq:first}
411: \begin{aligned}
412: ds^2&=(1+|g|^2)^2 \omega\cdot\bar \omega\;, \\
413: I\!I&=-Q-\overline Q + ds^2\;,
414: \end{aligned}
415: \end{equation}
416: where $Q=\omega\cdot dg$ is the Hopf differential of $F$.
417:
418: Let $f=FF^*\colon{}M\to H^3$ be a complete conformal \cmcone{}
419: immersion whose total Gaussian curvature is finite.
420: Since the Gaussian curvature $K$ of \cmcone{} surface is non-negative,
421: finiteness of the total Gaussian curvature is equivalent to
422: \[
423: \int_M (-K)\, dA < \infty\;,
424: \]
425: where $dA$ is the area element with respect to the first fundamental
426: form.
427: Then there is a compact Riemann surface $\overline M$ and a
428: finite number of points $\{p_1,$ $\dots,$ $p_n\}$ $\in \overline M$ such
429: that $M=\overline M\setminus \{p_1,\dots,p_n\}$.
430: We call each $p_j$ an {\it end\/} of $f$.
431: The hyperbolic Gauss map $G$ on $M$ does not necessarily extend
432: meromorphically on $\overline M$.
433: The end $p_j$ is called a {\it regular\/} end if $p_j$ is at most a pole
434: singularity of $G$, and otherwise is called an {\it irregular} end.
435: Namely, an irregular end is an end at which the hyperbolic Gauss map
436: has an essential singularity.
437:
438: On the other hand, the Hopf differential $Q$ can always be extended as a
439: meromorphic $2$-differential on $\overline M$.
440: We denote by $\ord_p Q$ the order of the first non-vanishing term
441: of the Laurent expansion of the Hopf differential $Q$ at
442: $p \in \overline M$.
443: (By definition, $\ord_p Q>0$ at zeros of $Q$ and $\ord_p Q<0$ at poles
444: of $Q$.)
445: The following lemma is well-known
446: (cf.~\cite{Bryant}, Lemma 2.3 of \cite{uy1}).
447: \begin{lemma}\label{lem:irreg}
448: An end $p_j$ is regular if and only if $\ord_{p_j} Q \geq -2$.
449: \end{lemma}
450: Now we set $d\sigma_f=(-K) ds^2$. Then $d\sigma^2_f$ is a pseudometric
451: of constant curvature $1$ with conical singularities
452: (see the appendix of this paper, and also Proposition~4 of \cite{Bryant}).
453: It follows from \eqref{eq:first} and the Gauss equation that
454: \begin{equation} \label{eq:g-metric}
455: d\sigma^2_f=\frac {4\,dg\cdot d\bar g}{(1+|g|^2)^2} \; .
456: \end{equation}
457: Hence $d \sigma^2_f$ is the pull-back of the Fubini-Study metric
458: $d\sigma^2_0$ on $\CP^1$ induced by the secondary Gauss map
459: $g\colon{}\widetilde{M}^2\to \C\cup \{\infty\} =\CP^1$.
460: By \eqref{eq:first} and \eqref{eq:g-metric} we have
461: \begin{equation}\label{eq:gauss}
462: ds^2\cdot d\sigma^2_f= 4\,Q\cdot\overline{Q} \; .
463: \end{equation}
464: In addition to having conical singularities at the ends $p_j$,
465: the pseudometric $d\sigma_f^2$ also has a conical singularity at each
466: umbilic point $q \in M$ of $f$. The conical order of $d\sigma_f^2$ at
467: each point is defined in the appendix of this paper.
468: Since $ds^2$ is positive definite at $q$, we have the following:
469: \begin{lemma}
470: At umbilic points, the conical order of $d\sigma^2_f$ equals the order
471: of $Q$.
472: \end{lemma}
473: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
474: \section{Reflections of an abstract spherical triangle}
475: \label{sec:reflection}
476:
477: In this section, we introduce a method to construct \cmcone{} surface
478: with irregular ends.
479: In \cite{ruy1}, examples with regular ends are constructed from the
480: holomorphic data $G$ (the hyperbolic Gauss map) and $Q$ (the Hopf
481: differential).
482: However, since the hyperbolic Gauss map has essential singularities
483: at irregular ends, it is hard to find an explicit expression of $G$ in
484: our case.
485: Thus, our construction is based on the secondary Gauss map $g$ and the
486: Hopf differential $Q$.
487: Though $g$ is not a well-defined meromorphic function on the surface,
488: the pseudometric $d\sigma^2_f$ as in \eqref{eq:g-metric} is a spherical
489: metric with conical singularities which is well-defined on the surface.
490: So, we start by constructing a spherical metric $d\sigma_f^2$ with
491: conical singularities, using reflections of spherical triangles (see
492: \cite{uy3}).
493:
494: \subsection*{Abstract spherical triangles}%
495: In this section, we consider abstract spherical triangles and their
496: extensions by reflection.
497: First, we shall define abstract triangles.
498: We set
499: \[
500: \Delta:=\{x+iy\in \C\,;\, x\ge 0,\,y\ge 0,\,x+y\le 1\}\;,
501: \]
502: and label each vertex $V_1$, $V_2$, $V_3$ of this closed triangular
503: region $\Delta$ as in Figure~\ref{fig:triangle}.
504: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
505: \begin{figure}
506: \setlength{\unitlength}{5mm}
507: \small
508: \begin{center}
509: \begin{tabular}{cc}
510: \begin{picture}(11,11)(-5,-5)
511: \put(-1,0){\vector(1,0){6}}
512: \put(0,-1){\vector(0,1){6}}
513: \put(0,4){\line(1,-1){4}}
514: \put(0.2,-0.2){\makebox(0,0)[lt]{$V_3$}}
515: \put(4,-0.2){\makebox(0,0)[lt]{$V_2$}}
516: \put(-0.2,4){\makebox(0,0)[rb]{$V_1$}}
517: \put(0,0){\arc{1.5}{-1.5707963}{0}}
518: \put(4,0){\arc{1.5}{-3.1415927}{-2.3561945}}
519: \put(0,4){\arc{1.5}{-5.4977871}{-4.712389}}
520: \put(0.8,0.8){\makebox(0,0)[cc]{$C$}}
521: \put(3.0,0.4){\makebox(0,0)[cc]{$B$}}
522: \put(0.4,3){\makebox(0,0)[cc]{$A$}}
523: \put(5.2,-0.4){\makebox(0,0)[cc]{$x$}}
524: \put(-0.1,5.3){\makebox(0,0)[cc]{$y$}}
525: \put(2,0){\vector(0,1){0.5}}
526: \put(2,0){\vector(0,-1){0.5}}
527: \put(0,2){\vector(1,0){0.5}}
528: \put(0,2){\vector(-1,0){0.5}}
529: \put(2,2){\vector(1,1){0.35355339}}
530: \put(2,2){\vector(-1,-1){0.35355339}}
531: \put(2.2,-0.3){\makebox(0,0)[lt]{$\mu_1$}}
532: \put(-0.3,1.8){\makebox(0,0)[rt]{$\mu_2$}}
533: \put(2.6,2){\makebox(0,0)[lt]{$\mu_3$}}
534: \end{picture}
535: &
536: \begin{picture}(11,11)(-5,-5)
537: \put(-5,0){\vector(1,0){10}}
538: \put(0,-5){\vector(0,1){10}}
539: \Thicklines
540: \put(0,4){\line(1,-1){4}}
541: \put(0,0){\line(1,0){4}}
542: \put(0,0){\line(0,1){4}}
543: \thinlines
544: \put(4,0){\line(0,1){4}}
545: \put(0,4){\line(1,0){4}}
546: \put(-4,0){\line(1,1){4}}
547: \put(0,-4){\line(1,1){4}}
548: \put(0.2,-0.2){\makebox(0,0)[lt]{$V_3$}}
549: \put(4,-0.2){\makebox(0,0)[lt]{$V_2$}}
550: \put(-0.2,4){\makebox(0,0)[rb]{$V_1$}}
551: \put(2,0){\vector(0,1){0.5}}
552: \put(2,0){\vector(0,-1){0.5}}
553: \put(0,2){\vector(1,0){0.5}}
554: \put(0,2){\vector(-1,0){0.5}}
555: \put(2,2){\vector(1,1){0.35355339}}
556: \put(2,2){\vector(-1,-1){0.35355339}}
557: \put(2.2,-0.3){\makebox(0,0)[lt]{$\mu_1$}}
558: \put(-0.3,1.8){\makebox(0,0)[rt]{$\mu_2$}}
559: \put(2.6,2){\makebox(0,0)[lt]{$\mu_3$}}
560: \put(1.2,1.2){\makebox(0,0)[cc]{$\Delta$}}
561: \put(-1.6,1.2){\makebox(0,0)[cc]{$\Delta_2$}}
562: \put(1.2,-1.6){\makebox(0,0)[cc]{$\Delta_1$}}
563: \put(3,3){\makebox(0,0)[cc]{$\Delta_3$}}
564: \end{picture}
565: \end{tabular}
566: \end{center}
567: \caption{The triangle $\Delta$ and its reflections.}
568: \label{fig:triangle}
569: \end{figure}
570: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
571: An {\it abstract spherical triangle\/} is a pair $(\Delta,d\sigma^2)$,
572: where $d\sigma^2$ is a Riemannian metric defined on $\Delta$ with
573: constant curvature $1$ such that the three edges forming the
574: boundary $\partial \Delta$ are geodesics.
575: Let $A$, $B$ and $C$ be the interior angles of $\Delta$ at the
576: vertices $V_1$, $V_2$, and $V_3$ with respect to the metric $d\sigma^2$
577: respectively.
578: The angles $A$, $B$ and $C$ are positive, but may take any positive
579: values, including those greater than or equal to $\pi$.
580: Moreover, if $A,B,C\not\in \pi \Z$, then we call the abstract spherical
581: triangle $(\Delta,d\sigma^2)$ {\it non-degenerate}.
582: The following fact is known:
583: \begin{lemma}[\cite{uy7}]
584: \label{lem:trig}
585: Let $(\Delta, d\sigma^2)$ be a non-degenerate abstract spherical
586: triangle, then the three angles $A$, $B$, $C$ satisfy the inequality
587: \begin{equation}\label{eq:tri}
588: \cos^2 A +\cos^2 B + \cos^2 C +
589: 2\cos A\cos B \cos C < 1\;.
590: \end{equation}
591: Conversely, if a triple of positive real numbers $(A,B,C)$ satisfies
592: \eqref{eq:tri}, then there exists a unique non-degenerate abstract
593: spherical triangle such that the angles at $V_1$, $V_2$ and $V_3$ are
594: $A$, $B$ and $C$ respectively.
595: \end{lemma}
596: \begin{proof}[Proof\/\footnotemark{}.]
597: \footnotetext{
598: Recently, an alternative proof and a geometric explanation of
599: this lemma were given in \cite{FH} and \cite{F}.
600: }
601: We take a double (two identical copies) of $(\Delta, d\sigma^2)$ and
602: glue them along their corresponding vertices and edges.
603: Then we get a conformal pseudometric on $S^2$ with three conical
604: singularities with conical angles $2A$, $2B$, $2C$.
605: Consequently, the conical orders are
606: \[
607: \frac{A-\pi}\pi \; ,\quad \frac{B-\pi}\pi \; ,\quad \frac{C-\pi}\pi
608: \]
609: respectively.
610: If $(\Delta, d\sigma^2)$ is non-degenerate, then $A,B,C\not \in \pi \Z$.
611: By Corollary~3.2 of \cite{uy7}, the metric on $S^2$ is irreducible.
612: Then \eqref{eq:tri} follows from (2.19) of \cite{uy7}.
613:
614: Conversely, suppose that \eqref{eq:tri} holds.
615: By Theorem 2.4 of \cite{uy7}, there exists a unique conformal
616: pseudometric on $S^2$ with three conical singularities with conical
617: angles $2A, 2B, 2C\not\in 2\pi \Z$.
618: The uniqueness of this pseudometric implies that it has a symmetry and
619: can be considered as a gluing of two identical non-degenerate spherical
620: triangles.
621: \end{proof}
622: The above lemma implies that a non-degenerate abstract spherical
623: triangle is uni\-quely determined by its angles $A,B,C$.
624: So we denote it by
625: \[
626: \Trig(A,B,C):=(\Delta,d\sigma^2).
627: \]
628: Now we fix a non-degenerate abstract spherical triangle
629: $\Trig(A,B,C)$.
630: Since $\Delta$ is simply connected, there exists a meromorphic function
631: $g\colon{}\Delta\to\C\cup\{\infty\}$ such that the pull-back of the
632: Fubini-study metric on $\C\cup\{\infty\}=\CP^1$ by $g$ is
633: $d\sigma^2$.
634: However, such a choice of the developing map has an ambiguity up to an
635: $\SU(2)$-matrix action $g\mapsto a\star g$ ($a\in\SU(2)$).
636: We shall now remove this ambiguity, using a normalization:
637: There exists a unique (up to sign) developing map
638: \[
639: g=g_{A,B,C}:\Delta\longrightarrow \C\cup\{\infty\}
640: \]
641: of $d\sigma^2$ satisfying (see Figure \ref{fig:triangle})
642: \begin{equation}\label{eq:normalize}
643: e^{-i C}g(V_1)\in \R\cup\{\infty\}\;,\quad
644: g(V_2)\in \R\cup\{\infty\}\quad
645: \text{and}\quad
646: g(V_3)=0\;.
647: \end{equation}
648: We call the developing map $g=g_{A,B,C}$ the
649: {\it normalized developing map\/} of the triangle $\Trig(A,B,C)$.
650:
651: Let $\mu_j$ ($j=1,2,3$) be the reflections of the triangle
652: $\Trig(A,B,C)$ across the three edges, as in Figure~\ref{fig:triangle}.
653: Let $\Delta_j$ be the closed domain obtained by reflecting $\Delta$ with
654: respect to $\mu_j$ (see Figure~\ref{fig:triangle}).
655: Then each reflection $\mu_j$ can be regarded as an involution on the
656: domain $\Delta\cup\Delta_j$.
657: \begin{lemma}[Monodromy principle]\label{lem:p-I}
658: Let $\Trig(A,B,C)$ be a non-degenerate abstract spherical triangle
659: and $g_{A,B,C}:\Delta \to \C\cup \{\infty\}$ $(j=1,2,3)$ be the
660: normalized developing map of $\Trig(A,B,C)$.
661: Then the following identities hold\/{\rm :}
662: \begin{align*}
663: \overline{ g_{A,B,C}\circ \mu_1}&=g_{A,B,C}\;,\\
664: \overline{ g_{A,B,C}\circ \mu_2}&=e^{-2iC}g_{A,B,C}\;,\\
665: \overline{ g_{A,B,C}\circ \mu_3}&=
666: \begin{pmatrix}
667: q & i \delta \\
668: i\delta & \bar q
669: \end{pmatrix} \star g_{A,B,C}
670: \qquad (\delta\in \R, \quad q\bar q+\delta^2=1)\;,
671: \end{align*}
672: where
673: \begin{equation}\label{eq:q-value}
674: q = \frac{i}{\sin C}(\cos A + e^{iC}\cos B)\;.
675: \end{equation}
676: \end{lemma}
677: \begin{proof}
678: To simplify the notation, we set $g=g_{A,B,C}$.
679: Let $\gamma_1$, $\gamma_2$ and $\gamma_3$ be the
680: three edges of $\Delta$
681: which are stabilized by the reflections $\mu_1$, $\mu_2$ and $\mu_3$,
682: respectively.
683: Since the edge $\gamma_1$ is a geodesic, \eqref{eq:normalize} implies
684: that $g(\gamma_1)$ lies on the real axis.
685: Then by the reflection principle, $\overline{g\circ\mu_1}=g$ holds.
686: Similarly, by \eqref{eq:normalize}, $e^{-iC}g(\gamma_2)$ is real.
687: Hence
688: \[
689: \overline{e^{-iC}g\circ\mu_2}=e^{-iC}g
690: \]
691: holds.
692: Then we have the second assertion.
693:
694: There exists a rotation $a$ centered at $g(V_2)$ of the unit 2-sphere
695: $S^2(=\C\cup\{\infty\})$ such that the image of $g(V_1)$ is real.
696: Such an isometry $a$ of $S^2$ can be represented as a matrix
697: $a\in \SU(2)$.
698: Then we have $a\star g(\gamma_3)$ lies on the real axis.
699: Hence by the reflection principle,
700: $\overline{a\star g\circ\mu_3} = a \star g$ holds, and then we have
701: \[
702: \overline{g\circ\mu_3} =
703: \left(\bar a^{-1}a\right)\star g\;.
704: \]
705: In particular, we have
706: \begin{equation}\label{eq:easy}
707: \overline{ g\circ \mu_1}=g,\quad
708: \overline{ g\circ \mu_2}=e^{-2iC}g,\quad
709: \overline{ g\circ \mu_3}=(\bar a^{-1}a)\star g\;.
710: \end{equation}
711: Now we glue $\Trig(A,B,C)$ to a double of $\Trig(A,B,C)$ along
712: corresponding edges and vertices, giving us
713: a constant curvature one metric of three conical singularities
714: on $S^2$ with conical angles $2A$, $2B$, and $2C$ just as
715: in the proof of Lemma \ref{lem:trig}.
716: The open domains $\Delta_1$, $\Delta_2$ and $\Delta_3$ can be regarded
717: as the interior of the second triangle in the double
718: of $\Trig(A,B,C)$.
719: The monodromy of reflections for such metrics on $S^2$ are determined
720: in \cite{uy7}.
721: Then, as shown at the bottom of page~82 of \cite{uy7}, we have
722: \[
723: \bar a^{-1}a=
724: \begin{pmatrix}
725: q & i \delta \\
726: i\delta & \bar q
727: \end{pmatrix},
728: \qquad (\delta\in \R,\, q\bar q+\delta^2=1)
729: \]
730: with $q$ as in \eqref{eq:q-value}.
731: \end{proof}
732: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
733: \subsection*{Closed Riemann surfaces generated by three reflections}
734: Let $\overline M$ be
735: a closed Riemann surface and $D(\subset \overline{M})$
736: a simply connected closed domain bounded by three real analytic curves
737: $\gamma_1$, $\gamma_2$ and $\gamma_3$.
738: We label the vertices $V_1$, $V_2$, $V_3$ of this closed
739: triangular region
740: $D$ such that $\gamma_1$, $\gamma_2$, and $\gamma_3$ correspond to the
741: three edges $V_2V_3$, $V_3V_1$ and $V_1V_2$, respectively.
742: The Riemann surface $\overline M$ is generated by $D$ if there are three
743: anti-holomorphic reflections $\mu_1$, $\mu_2$, $\mu_3$ of $\overline M$
744: stabilizing the three edges $\gamma_1$, $\gamma_2$, and $\gamma_3$ of
745: $D$ such that any point of
746: $\overline M$ can be contained in the image of $D$ by a suitable finite
747: composition of these three reflections (see Figure~\ref{fig:domain}).
748: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
749: \begin{figure}
750: \begin{center}
751: \includegraphics[width=4cm]{figure2.eps}
752: \end{center}
753: \caption{A Riemann surface generated by reflections:
754: Here the Riemann surface is the sphere $S^2$, and is obtained
755: from the triangle $D$ by reflections.}
756: \label{fig:domain}
757: \end{figure}
758: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
759: In this case, $D$ is called a {\it fundamental domain\/} of
760: $\overline M$.
761: A meromorphic $2$-differential $Q$ on $\overline M$ is said to be
762: symmetric with respect to $D$ if
763: \[
764: \overline{Q\circ \mu_j}=Q\qquad (j=1,2,3)
765: \]
766: holds.
767:
768: We let $\metone(\overline M)$ denote the set of conformal pseudometrics
769: with conical singularities on $\overline M$.
770: A metric $d\sigma^2 \in \metone(\overline M)$ is called
771: {\it symmetric\/} with respect to the fundamental domain
772: $D$ if it is invariant under the three
773: reflections $\mu_1$, $\mu_2$ and $\mu_3$.
774: Moreover, if the restriction $(D, d\sigma^2|_{D})$ is isometric to
775: $\Trig(A,B,C)$, we denote the metric by
776: \[
777: d\sigma^2=d\sigma^2_{A,B,C}\;.
778: \]
779:
780: A meromorphic function $g$ on $\overline M$ is called
781: {\it $\SU(2)$-symmetric\/} with respect to $D$ if
782: \begin{equation}\label{eq:dsigmag}
783: d\sigma^2_g :=\frac{4\,dg\cdot d\bar g}{(1+|g|^2)^2}
784: \end{equation}
785: is symmetric with respect to $D$.
786:
787: The following is the main theorem in this paper:
788: \begin{theorem}\label{thm:gen}
789: Let $\overline M$ be a closed Riemann surface which is generated
790: by a triangular fundamental domain $D\subset\overline M$ by
791: reflections, and label the vertices of $D$ as $V_1$, $V_2$ and $V_3$.
792: Let $g$ be an $\SU(2)$-symmetric meromorphic function on $\overline M$
793: with respect to $D$, and let $Q$ be a symmetric meromorphic
794: $2$-differential on $\overline M$ with respect to $D$.
795: Suppose that{\rm :}
796: \begin{enumerate}
797: \item\label{item:gen:angle}
798: There exist $A,B_0 \in \R^+ \setminus \pi \Z$ such that
799: $d\sigma^2_g=d\sigma^2_{A,B_0,\pi/2}$,
800: with $d\sigma^2_g$ as in \eqref{eq:dsigmag}.
801: \item\label{item:gen:ord}
802: $Q$ is holomorphic on $D\setminus \{V_2\}$ and has a pole at
803: $V_2$ with $\ord_{V_2}Q\leq -3$.
804: \item\label{item:gen:branch}
805: The branch point set of $g$ outside the poles of $Q$
806: equals to the zero set of $Q$, and at each zero of $Q$,
807: the order of $Q$ is equal to the conical order of $d\sigma^2_g$.
808: \setcounter{tempcounteri}{\value{enumi}}
809: \end{enumerate}
810: Let $p_1,\dots p_n \in \overline{M}$ be the set of poles of $Q$.
811: Then, for some $\varepsilon>0$, there exist a
812: smooth function $B(t)\colon{}(-\varepsilon,\varepsilon)\to\R$
813: satisfying $B(0)=B_0$
814: and a $1$-parameter family of conformal \cmcone{} immersions
815: $f_t:\overline M\setminus\{p_1,\dots,p_n\} \to
816: H^3$ for $t\in (-\varepsilon,\varepsilon)$ with the following
817: properties{\rm :}
818: \begin{enumeraterom}
819: \item\label{item:gen:concl-1}
820: The Hopf differential of $f_t$ is $tQ$ and
821: $d\sigma^2_{f_t}=d\sigma_{A,B(t),\pi/2}$.
822: \item\label{item:gen:concl-3}
823: $f_t$ has irregular ends at $\{p_1,\dots,p_n\}$.
824: \item\label{item:gen-concl-4}
825: $f_t$ is symmetric with respect to $D$.
826: That is, the image of $f_t$ is generated by the reflections
827: across the edges of $f_t(D)$.
828: \setcounter{tempcounterii}{\value{enumi}}
829: \end{enumeraterom}
830: \end{theorem}
831: \begin{remark}
832: The construction method for proving Theorem~\ref{thm:gen} will
833: still work if $\ord_{V_2}Q=-2$. %
834: The stronger assumption $\ord_{V_2}Q\leq -3$ in \ref{item:gen:ord}
835: is required only to make the ends irregular
836: (see Lemma~\ref{lem:irreg}).
837: \end{remark}
838: The next theorem gives conditions which imply the surfaces $f_t$ in
839: Theorem~\ref{thm:gen} are irreducible.
840: \begin{theorem}
841: \label{thm:irred}
842: Under the assumptions in Theorem~{\rm \ref{thm:gen}}, if
843: $A\not\equiv \pi/2 \pmod{\pi}$ and
844: \begin{equation}\label{eq:irred-cond}
845: \oint_{\tau}
846: \begin{pmatrix}
847: g & - g^2 \\
848: 1 & -g\hphantom{^2}
849: \end{pmatrix}\frac{Q}{dg}\neq 0\;
850: \end{equation}
851: for a local loop $\tau$ about $V_2$, then
852: $f_t$ is irreducible for $t$ sufficiently close to zero.
853: \end{theorem}
854: The proofs of these theorems are given in Section~\ref{sec:proof}.
855: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
856: \section{CMC-1 surfaces with dihedral and Platonic symmetries}
857: \label{sec:platonic}
858: In this section, we construct examples of finite total curvature
859: \cmcone{} surfaces with irregular ends and either dihedral or Platonic
860: symmetries.
861: Examples with dihedral symmetries and the simplest example with
862: tetrahedral symmetry (the case $m=1$ in \eqref{eq:tetraQ} and
863: \eqref{eq:tetrag}) are irreducible.
864: (Though we expect all the other examples to be irreducible, we have not
865: checked them yet.)
866:
867: \subsection*{CMC-1 surfaces with dihedral symmetries}
868: Let $n\geq 3$ be an integer and
869: \[
870: M_n:=\C\cup\{\infty\}\setminus\{1,\zeta,\dots,\zeta^{n-1}\}
871: \qquad \left(\zeta=\exp\frac{2\pi}{n} i\right)\;.
872: \]
873: The Jorge-Meeks' $n$-noid is the complete minimal immersion
874: $f_{n,0}\colon{}M_n\to\R^3$ given by the Weierstrass representation
875: as
876: \begin{multline*}
877: f_{n,0}:= \Re\int
878: \bigl((1-g_{n,0}^2), i(1+g_{n,0}^2),2 g_{n,0}\bigr)
879: \frac{Q_{n,0}}{dg_{n,0}}\;,\\
880: \text{where}\qquad
881: g_{n,0} = z^{n-1},\quad
882: Q_{n,0} = \frac{z^{n-2}}{(z^n-1)^2}\,dz^2\;.
883: \end{multline*}
884: The $\Z_2$ extension $D_n \times \Z_2$ of the dihedral group
885: $D_n$ acts isometrically on the image of $f_{n,0}$.
886:
887: There exists a one-parameter family of corresponding
888: \cmcone{} immersions of $M_n$ to $H^3$ such that the
889: hyperbolic Gauss map is $g_{n,0}$ and the Hopf differential is
890: proportional to $Q_{n,0}$ (see \cite{uy3,ruy1}, and
891: Figure~\ref{fig:trinoid} for the $n=3$ case).
892: Since $\ord_{\zeta^j}Q_{n,0}=-2$ $(j=0,\dots,n-1)$, the ends of
893: these corresponding surfaces are regular.
894: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
895: \begin{figure}
896: \setlength{\unitlength}{2cm}
897: \small
898: \begin{center}
899: \begin{picture}(3,3)(-1.5,-1.5)
900: \put(-1.5,0){\vector(1,0){3}}
901: \put(0,-1.5){\vector(0,1){3}}
902: \put(0,0){\circle{2}}
903: \put(1,0){\circle*{0.05}}
904: \put(0.30901699,0.95105652){\circle*{0.05}}
905: \put(0.30901699,-0.95105652){\circle*{0.05}}
906: \put(-0.80901699,0.58778525){\circle*{0.05}}
907: \put(-0.80901699,-0.58778525){\circle*{0.05}}
908: \drawline(0,0)(1.2,0)
909: \drawline(0,0)(0.30901699,0.95105652)
910: \drawline(0,0)(0.80901699,0.58778525)
911: \put(1.1,-0.1){\makebox(0,0){$1$}}
912: \put(0.4,1.1){\makebox(0,0){$\zeta$}}
913: \put(-0.1,-0.1){\makebox(0,0){$0$}}
914: \put(0.6,0.2){\makebox(0,0){\normalsize $\Omega_n$}}
915: \put(1.18,0.65){\makebox(0,0){$V_3=e^{i \pi/n}$}}
916: \put(0.1,-0.1){\makebox(0,0){$V_1$}}
917: \put(0.9,-0.1){\makebox(0,0){$V_2$}}
918: \end{picture}
919: \end{center}
920: \caption{The fundamental domain of surfaces with dihedral symmetry}
921: \label{fig:fund-jm}
922: \end{figure}
923: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
924:
925: However, as we wish to produce \cmcone{} surfaces in $H^3$ whose
926: ends are {\em not} regular,
927: we now modify our choices for $Q$ and $g$ to accomplish this:
928: Let $m\geq 1$ be an integer and set
929: \[
930: Q_{n,m}:= \frac{z^{n(m+1)-2}}{(z^n-1)^{2(m+1)}}\,dz^2,\qquad
931: g_{n,m}:= z^{n(m+1)-1}\;.
932: \]
933: Then $g=g_{n,m}$ and $Q=Q_{n,m}$ are a meromorphic function and
934: a meromorphic $2$-differential on $\overline M=\C\cup\{\infty\}$
935: respectively,
936: which are symmetric with respect to the fundamental domain $\Omega_n$ as
937: in Figure~\ref{fig:fund-jm}.
938: Moreover, $(g,Q)$ satisfies the assumptions
939: \ref{item:gen:angle}--\ref{item:gen:branch} of
940: Theorem~\ref{thm:gen} for $A=\pi(m+1)-\pi/n$
941: and $B_0=\pi/2$.
942: Then, for each $n$ and $m$, there exists a one-parameter family of
943: \cmcone{} immersions $f_{n,m,t}\colon{}M_n\to H^3$ ($0<|t|<\varepsilon$)
944: with symmetry group
945: $D_n \times \Z_2$ whose Hopf differential is $tQ_{n,m}$
946: (see Figure~\ref{fig:trinoid}). The total Gaussian curvature of these
947: surfaces will be approximately $4\pi (n(m+1)-1)$.
948: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
949: \begin{figure}
950: \small
951: \begin{center}
952: \begin{tabular}{c@{\hspace{1cm}}c}
953: \includegraphics[width=3cm]{figure4a.eps} &
954: \includegraphics[width=3cm]{figure4b.eps} \\
955: (a) $f_{3,0,t}$ &
956: (b) Fundamental piece of $f_{3,0,t}$ \\[2ex]
957: \includegraphics[width=3cm]{figure4c.eps} &
958: \includegraphics[width=3cm]{figure4d.eps} \\
959: (c) Fundamental piece of $f_{3,1,t}$ &
960: (d) Enneper's cousin
961: \end{tabular}
962: \end{center}
963: \caption{Surfaces with dihedral symmetry:
964: Figure (a) shows a \cmcone{} surface with
965: dihedral symmetry and
966: three regular ends (the surface corresponding to
967: a Jorge-Meeks surface) in the Poincar\'e model of $H^3$,
968: and figure (b) is the fundamental region of the surface in
969: figure (a).
970: Figure (c) shows the fundamental piece of $f_{3,1,t}$,
971: a surface of dihedral symmetry with three irregular ends.
972: The central part of $f_{3,1,t}$ is similar to that of the
973: hyperbolic correspondence of a Jorge-Meeks surface.
974: The end of $f_{3,1,t}$ seen here is similar to the end of an
975: Enneper cousin \cite{Bryant}, which is shown in figure (d).}
976: \label{fig:trinoid}
977: \end{figure}
978: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
979:
980: Let $\tau$ be a loop surrounding the end $V_2=1$. Since
981: \[
982: \frac{Q_{n,m}}{dg_{n,m}}
983: =\left(\frac{1}{n(m+1)-1}\right)\frac{dz}{(z^n-1)^{2(m+1)}}\;,
984: \]
985: we have
986: \[
987: \oint_{\tau}\frac{Q_{n,m}}{dg_{n,m}}=
988: \left(\frac{2\pi i}{n(m+1)-1}\right)
989: \Res_{z=1} \frac{1}{(z^n-1)^{2(m+1)}} \neq 0\;.
990: \]
991: Thus by Theorem~\ref{thm:irred}, the surfaces are irreducible
992: for sufficient small $t$.
993: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
994: \subsection*{CMC-1 surfaces with tetrahedral symmetries}
995: It is well-known that there exists a minimal immersion
996: \[
997: f_{0}\colon{}M
998: :=\C\cup\{\infty\}\setminus\{p_1,\dots,p_4\}
999: \longrightarrow
1000: \R^3
1001: \]
1002: with $4$ catenoid ends and tetrahedral symmetry \cite{Kat,Xu,BR,uy3}
1003: and corresponding \cmcone{} surfaces in $H^3$ \cite{uy3,ruy1} with
1004: regular ends.
1005:
1006: We denote by $Q_{0}$ and $g_{0}$ the Hopf differential and the
1007: Gauss map of $f_{0}$ respectively.
1008: Since each end is asymptotic to a catenoid, $\ord_{p_j}Q_{0}=-2$
1009: ($j=1,\dots,4$).
1010: Then there exists $4$ umbilic points (zeros of $Q_{0}$)
1011: $q_1,\dots,q_4$ such that $\ord_{q_j}Q_{0}=1$ ($j=1,\dots,4$).
1012: The Gauss map $g_{0}$ is a meromorphic function on
1013: $\C\cup\{\infty\}$ whose branch points are $\{q_1,\dots,q_4\}$
1014: each with branch order $1$.
1015: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
1016: \begin{figure}
1017: \setlength{\unitlength}{7mm}
1018: \small
1019: \begin{center}
1020: \begin{picture}(7,7)
1021: \Thicklines
1022: \drawline(4,6)(4.5,1)(1,2)(4,6)
1023: \drawline(4,6)(6,3)(4.5,1)
1024: \thinlines
1025: \dottedline{0.1}(1,2)(6,3)
1026: \drawline(4,6)(2.75,1.5)
1027: \drawline(4.5,1)(2.5,4)
1028: \drawline(1,2)(4.25,3.5)
1029: \put(4,6.4){\makebox(0,0){$p_1$}}
1030: \put(4.5,0.6){\makebox(0,0)[cc]{$p_2$}}
1031: \put(0.6,2){\makebox(0,0)[cc]{$p_3$}}
1032: \put(6.4,3){\makebox(0,0)[cc]{$p_4$}}
1033: \put(2.7,3.15){\makebox(0,0)[cc]{$q_1$}}
1034: \put(3.8,4){\makebox(0,0)[cc]{\normalsize$D$}}
1035: \end{picture}
1036: \end{center}
1037: \caption{The fundamental domain of surfaces with tetrahedral symmetry}
1038: \label{fig:fund-tet}
1039: \end{figure}
1040: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
1041: Moreover, $M$ is obtained by reflections across the edges of the
1042: fundamental domain $D$, which is a triangle with vertices
1043: $V_1=q_1$, $V_2=p_1$, $V_3$. (See Figure~\ref{fig:fund-tet}.
1044: See also the construction in \cite{uy3}.)
1045: The Hopf differential $Q_{0}$ and the Gauss map $g_{0}$ are symmetric
1046: with respect to the fundamental domain $D$.
1047:
1048: Consider the Schwarzian derivative $S(g_{0})$ of $g_{0}$, as in
1049: \eqref{eq:sch},
1050: where $z$ is the usual complex coordinate of $\C\cup\{\infty\}$.
1051: Then $S(g_0)$ is a meromorphic $2$-differential on $\C\cup\{\infty\}$.
1052: Moreover, since $g_{0}$ is symmetric, $S(g_0)$ is invariant under
1053: reflections about the edges of $D$.
1054:
1055: The branch points of $g_0$ are poles of $S(g_0)$,
1056: and each pole of $S(g_{0})$ has order $-2$.
1057: So $S(g_0)$ has $4$ poles of order $2$ at the $q_j$ ($j=1,\dots,4$) and
1058: is holomorphic on $\C\cup\{\infty\}\setminus\{q_1,\dots,q_4\}$.
1059: Since the total order of a meromorphic $2$-differential on
1060: $\C\cup\{\infty\}$ is $-4$, $S(g_0)$ has $4$ zeros (counting
1061: multiplicity).
1062: If there exists a zero of $S(g_0)$ on the interior of $D$, $S(g_0)$ has
1063: at least $24$ zeros because $\C\cup\{\infty\}$ consists of $24$ copies
1064: of the fundamental region $D$, which is impossible.
1065: Similarly, if a zero of $S(g_0)$ lies on the interior of a edge of $D$,
1066: $S(g_0)$ has at least $12$ zeros, which is also impossible.
1067: If the vertex $V_3$ of $D$ is a zero of $S(g_0)$, there exist $6$ zeros,
1068: which is again impossible.
1069: Since $V_1=q_1$ is a pole of $S(g_0)$, the set of zeros of $S(g_0)$ must
1070: be $\{p_1,\dots,p_4\}$, and $\ord_{p_j}S(g_0)=1$ for $j=1,\dots 4$.
1071:
1072: For an integer $m\geq 1$, we define a meromorphic $2$-differential
1073: $Q_{m}$ as
1074: \begin{equation}\label{eq:tetraQ}
1075: Q_{m}:=\frac{Q_{0}{}^{m+1}}{S(g_0)^m}\;,
1076: \end{equation}
1077: where $Q_{0}{}^{m+1}$ (resp.~$S(g_0)^m$) is the symmetric product of
1078: $m+1$ copies of $Q_{0}$ (resp.~$m$-copies of $S(g_0)$).
1079: Since the poles of $Q_{0}$ and the zeros of $S(g_0)$ are
1080: $\{p_1,\dots,p_4\}$,
1081: $Q_{m}$ has the poles $\{p_1,\dots,p_4\}$ and is holomorphic on $M$.
1082: More precisely,
1083: \[
1084: \ord_{p_j}Q_{m}= -3m-2\qquad\text{and}\qquad
1085: \ord_{q_j}Q_{m}= 3m+1\qquad (j=1,\dots,4)
1086: \]
1087: hold.
1088: Since $Q_{0}$ and $S(g_0)$ are invariant under the reflections,
1089: so is $Q_{m}$.
1090:
1091: Consider an abstract spherical triangle $\Trig(A,B_0,C)$ with
1092: \begin{equation}\label{eq:tetrag}
1093: A = \pi m + \frac{2}{3}\pi\;,\quad
1094: B_0 = \frac{\pi}{3}\;,\quad
1095: C = \frac{\pi}{2}\;,
1096: \end{equation}
1097: and identify it with the fundamental domain $D$.
1098: Then we have a pseudometric
1099: $d\sigma^2_{A,B_0,C}\in\metone(\C\cup\{\infty\})$.
1100: Since neighborhoods of $V_1$, $V_2$ and $V_3$ are generated by
1101: $6$, $6$ and $4$ copies of the fundamental domain $D$, respectively,
1102: $d\sigma^2_{A,B_0,C}$ is a pseudometric whose conical orders are all
1103: integers.
1104: Hence by Proposition~\ref{prop:A} in Appendix, the developing map
1105: $g_{m}$ of $d\sigma^2_{A,B_0,C}$ is meromorphic on $\C\cup\{\infty\}$.
1106:
1107: Then one can easily check that $(g_{m},Q_{m})$ satisfies
1108: the assumptions \ref{item:gen:angle}--\ref{item:gen:branch} of
1109: Theorem \ref{thm:gen}.
1110: Hence for each $m$, there exists a one-parameter family of \cmcone{}
1111: immersions $\{f_{m,t}\,;\,0<|t|<\varepsilon\}$ of $M$ into $H^3$ with
1112: irregular ends and tetrahedral symmetry.
1113:
1114: Finally, we check irreducibility for $m=1$.
1115: We may set
1116: \[
1117: M = \C\cup\{\infty\}\setminus\{1,\zeta,\zeta^2,\infty\}\;,
1118: \qquad\text{where}\quad \zeta =\exp\frac{2}{3}\pi i\;,
1119: \]
1120: and
1121: \[
1122: g_0 = \frac{1}{3\sqrt{2}}\left(z-\frac{4}{z^3}\right),\qquad
1123: Q_0 = \frac{z(z^3+8)}{(z^3-1)^2}\,dz^2
1124: \]
1125: (see page 221 of \cite{uy3}).
1126: Hence the umbilic points are $\{0,-2,-2\zeta,-2\zeta^2\}$.
1127: By direct calculation, we have
1128: \[
1129: Q_1 = \frac{Q_0{^2}}{S(g_0)}=
1130: \frac{1}{96}\frac{z^4(z^3+8)^4}{(z^3-1)^5}\,dz^2\;.
1131: \]
1132: On the other hand, $g_{1}$ is a meromorphic function which branches at
1133: the umbilic points with branch order $4$.
1134: Then we have $\deg g_1=9$ by the Riemann-Hurwitz formula.
1135: Choose a rotation $a\in\SU(2)$ such that $a\star g_1(q_1)=\infty$,
1136: where we set $q_1=0$.
1137: Then $q_j$ $(j=2,3,4)$ are not poles of $g:=a\star g_1$,
1138: because the multiplicity of $g$ at $q_j$ is $5$ for each $j$
1139: and $\deg g=9$.
1140: Moreover, $d\sigma^2_{g_1}=d\sigma^2_g$ has a conical singularity at $0$
1141: with conical order $4$.
1142: Hence, by symmetricity, we have $g_1(\zeta z)=\zeta g_1(z)$, and we may write
1143: \[
1144: dg = \beta \frac{(z^3+8)^4}{z^6(z^3-a^3)^2}\,dz
1145: \]
1146: for some nonzero constants $a$ and $\beta$.
1147: Such a function $g$ exists if and only if all residues of the right-hand
1148: side vanish, which is equivalent to $a^3=16$.
1149: Then one can check irreducibility by direct calculation and
1150: Theorem~\ref{thm:irred}.
1151: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
1152: \subsection*{CMC-1 surfaces with Platonic symmetries}
1153: There are genus zero minimal surfaces in $\R^3$ with catenoidal ends and
1154: the symmetry of any Platonic solid (\cite{Kat,Xu,BR,uy3}).
1155: By similar arguments to the tetrahedral case above, one can obtain
1156: \cmcone{}
1157: immersions with irregular ends and any Platonic symmetry.
1158: Table~\ref{tab:platonic} shows the data for such surfaces.
1159: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
1160: \begin{table}
1161: \footnotesize
1162: \begin{center}
1163: \newcommand{\vr}{\rule[-1.5ex]{0mm}{2.8ex}}
1164: \begin{tabular}{|c||c|c|c|c|c|c|c|}
1165: \hline
1166: Symmetry &
1167: {$\# \{p_j\}$}&
1168: {$\# \{q_j\}$}&
1169: {$\ord_{q_j}Q_0$} &
1170: {$\ord_{p_j}Q_m$} &
1171: {$\ord_{q_j}Q_m$} &
1172: {$A$} &
1173: {$B_0$} \\
1174: \hline
1175: \hline
1176: Dihedral &
1177: $n$ & $2$ & $n-2$ & $2(m+1)$ & $n(m+1)-2$ &
1178: $\pi\left(m+1-\frac{1}{n}\right)$ & $\frac{\pi}{2}$\vr\\
1179: \hline
1180: Tetrahedral &
1181: $4$ & $4$ & $1$ & $-3m-2$ & $3m+1$ & $\pi\left(m+\frac{2}{3}\right)$ &
1182: $\frac{\pi}{3}$\vr\\
1183: \hline
1184: Octahedral &
1185: $8$ & $6$ & $2$ & $-3m-2$ & $4m+2$ & $\pi\left(m+\frac{4}{3}\right)$ &
1186: $\frac{\pi}{3}$\vr\\
1187: \hline
1188: Octahedral &
1189: $6$ & $8$ & $1$ & $-4m-2$ & $3m+1$ & $\pi\left(m+\frac{2}{3}\right)$ &
1190: $\frac{\pi}{4}$\vr\\
1191: \hline
1192: Icosahedral &
1193: $20$ & $12$ & $3$ & $-3m-2$ & $5m+3$ & $\pi\left(m+\frac{4}{5}\right)$
1194: & $\frac{\pi}{3}$\vr\\
1195: \hline
1196: Icosahedral &
1197: $12$ & $20$ & $1$ & $-5m-2$ & $3m+1$ & $\pi\left(m+\frac{2}{3}\right)$
1198: & $\frac{\pi}{5}$\vr\\
1199: \hline
1200: \end{tabular}
1201: \end{center}
1202: \caption{Data for \cmcone{} surfaces with Platonic symmetries}
1203: \label{tab:platonic}
1204: \end{table}
1205: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
1206: \section{Proof of the main theorem}
1207: \label{sec:proof}
1208: \subsubsection*{Proof of Theorem~\ref{thm:gen}}%
1209: \newcommand{\tempskip}{\vspace{0.5ex}}
1210: \begin{proof}[Step 1{\rm :}]
1211: Take a real number $B\not\in \pi\Z$ and let
1212: \begin{equation}\label{eq:rho}
1213: \begin{aligned}
1214: \rho_1&=\id,\qquad
1215: \rho_2=\begin{pmatrix}
1216: -i & 0 \\
1217: \hphantom{-}0 & i
1218: \end{pmatrix} \; ,\\
1219: \rho_3&=\rho_3(B)=
1220: \begin{pmatrix}
1221: q(B) & i \delta(B) \\
1222: i\delta(B) & \bar q(B)
1223: \end{pmatrix}
1224: \qquad (\delta\in \R,~ q\bar q+\delta^2=1) \; ,
1225: \end{aligned}
1226: \end{equation}
1227: where
1228: \begin{equation}\label{eq:C}
1229: q(B) =i\cos A -\cos B \; .
1230: \end{equation}
1231: Then we have
1232: \[
1233: \bar \rho_j \rho_j=\id\qquad (j=1,2,3) \; .
1234: \]
1235: Since $\Trig(A,B_0,\pi/2)$ is non-degenerate (by the assumption
1236: \ref{item:gen:angle}),
1237: $A$, $B_0$ and $C=\pi/2$ satisfy the relation \eqref{eq:tri}.
1238: Then for each $B$ sufficiently close to $B_0$,
1239: there exists an abstract spherical triangle with angles $A$, $B$ and
1240: $\pi/2$.
1241: We identify the domain $D\subset\overline M$ with this triangle.
1242: Then by reflecting the metric,
1243: we have $d\sigma^2_{A,B,\pi/2}\in\metone(\overline M)$.
1244: Let $M:=\overline M\setminus\{p_1,\dots,p_n\}$ and
1245: $\pi\colon{}\widetilde M\to M$ the universal covering.
1246: By Proposition~\ref{prop:A} in the appendix
1247: of this paper, the developing map
1248: $\hat g_{A,B,\pi/2}$ of $d\sigma^2_{A,B,\pi/2}$ is defined on
1249: $\widetilde M$.
1250: To simplify the notation, we set
1251: \[
1252: \hat g_B:=\hat g_{A,B,\pi/2}\colon{}
1253: \widetilde M \longrightarrow \C\cup \{\infty\} \; .
1254: \]
1255: Then by the monodromy principle (Lemma~\ref{lem:p-I}), we have
1256: \[
1257: \overline{\hat g_B\circ \mu_j}=\rho_j\star \hat g_B \qquad (j=1,2,3) \; .
1258: \]
1259: \renewcommand{\qedsymbol}{}
1260: \end{proof}
1261: \begin{proof}[Step 2{\rm :}]
1262: One may regard the triangle $D\setminus \{V_2\}$ as generating
1263: $\widetilde M$ by the three reflections across its edges.
1264: We denote these reflections
1265: by $\hat \mu_1$, $\hat \mu_2$, and $\hat \mu_3$;
1266: that is, each $\hat \mu_j$ ($j=1,2,3$) is an antiholomorphic
1267: transformation on $\widetilde M$ which preserves the $j$'th edge of
1268: the triangle $D\setminus \{V_2\}$.
1269: We set
1270: \[
1271: \hat Q:=Q\circ \pi \;.
1272: \]
1273: Let $F=F_{t,B}$ be a solution of the following ordinary differential
1274: equation on $\widetilde M$:
1275: \begin{equation}\label{eq:our_ode}
1276: F^{-1}dF=
1277: t\begin{pmatrix}
1278: {\hat g}_B & -{\hat g}_B^2 \\
1279: 1 & -{\hat g}_B
1280: \end{pmatrix} \frac{\hat Q}{d{\hat g}_B},\qquad F(V_3)=\id\;.
1281: \end{equation}
1282: Such a solution $F_{t,B}$ is uniquely determined on $\widetilde M$.
1283: By \eqref{eq:our_ode}, the right-hand side of the ordinary
1284: differential equation is traceless and so $F_{t,B}$ takes values in
1285: $\SL(2,\C)$.
1286:
1287: Then $\overline{F_{t,B}\circ \hat\mu_j}$ ($j=1,2,3$) has the Hopf
1288: differential $\hat Q=\overline{\hat Q\circ \hat \mu_j}$ and the secondary
1289: Gauss map satisfies
1290: $\overline{\hat g_B\circ \hat\mu_j}=\rho_j\star \hat g_B$.
1291: However, $F_{t,B}\,\rho_j^{-1}$ also has the Hopf differential
1292: $\hat Q$ and secondary Gauss map $\rho_j\star \hat g_B$, by
1293: \eqref{eq:three}.
1294: Thus, by \eqref{eq:our_ode}, we have
1295: \[
1296: \left(\overline{F_{t,B}\circ \hat\mu_j}\right)^{-1}\!\!
1297: d \left(\overline{F_{t,B}\circ \hat\mu_j}\right)=
1298: \left(F_{t,B}\,\rho_j^{-1}\right)^{-1}
1299: d\left(F_{t,B}\,\rho_j^{-1}\right)=
1300: \rho_j
1301: \begin{pmatrix}
1302: \hat g_B & -\hat g_B^2 \\
1303: 1 & -\hat g_B
1304: \end{pmatrix} \frac{\hat Q}{d\hat g_B}\,\rho_j^{-1}\;.
1305: \]
1306: This implies that $\overline{F_{t,B}\circ \hat\mu_j}$ and
1307: $F_{t,B}\,\rho_j^{-1}$ are both solutions of the same ordinary
1308: differential equation, and thus they differ only by the choice of
1309: initial values at the base point $V_3$.
1310: So there exists a matrix $\sigma_j(t,B)\in \SL(2,\C)$ such that
1311: \begin{equation}\label{eq:sigma}
1312: \overline{F_{t,B}\circ \hat\mu_j}=\sigma_j(t,B)\,
1313: F_{t,B}\, \rho_j^{-1}\qquad (j=1,2,3).
1314: \end{equation}
1315: Since $\overline{F_{t,B}\circ \hat\mu_j(V_3)}=F(V_3)=\id$
1316: for $j=1,2$,
1317: \[
1318: \id=\sigma_j(t,B)\,\rho_j^{-1}\qquad (j=1,2)\;.
1319: \]
1320: holds.
1321: Thus we have
1322: \begin{equation}\label{eq:A}
1323: \sigma_1(t,B)=\rho_1=\id,
1324: \qquad \sigma_2(t,B)=\rho_2=
1325: \begin{pmatrix}
1326: -i & 0 \\
1327: \hphantom{-}0 & i
1328: \end{pmatrix}\;.
1329: \end{equation}
1330: In particular, the matrices $\sigma_1(t,B)$ and $\sigma_2(t,B)$ do not
1331: depend on $t$ nor the angle $B$.
1332: Since $F_{0,B}=\id$, \eqref{eq:sigma} implies that
1333: \begin{equation}\label{eq:q}
1334: \sigma_3(0,B)=\rho_3(B)
1335: =\begin{pmatrix}
1336: q(B) & i \delta(B) \\
1337: i\delta(B) & \overline{q(B)}
1338: \end{pmatrix}\;.
1339: \end{equation}
1340: \renewcommand{\qedsymbol}{}
1341: \end{proof}
1342: \begin{proof}[Step 3{\rm :}]
1343: Next we shall describe the matrix $\sigma_3(t,B)$.
1344: We have
1345: \begin{align*}
1346: F_{t,B}&=F_{t,B}\circ \hat\mu_3 \circ \hat\mu_3
1347: =\overline{%
1348: \overline{F_{t,B}\circ \hat\mu_3 \circ \hat\mu_3}}\\
1349: &=\overline{%
1350: \sigma_3(t,B)\bigl(F_{t,B}\circ \hat\mu_3\bigr)
1351: \rho_3(B)^{-1}}
1352: =\overline{%
1353: \sigma_3(t,B)}\,\sigma_3(t,B)\,F_{t,B}\,\rho_3(B)^{-1}
1354: \overline{\rho_3(B)^{-1}}\\
1355: &=\overline{\sigma_3(t,B)}\,\sigma_3(t,B)\,F_{t,B}\;,
1356: \end{align*}
1357: where we used the fact $\overline{\rho_3(B)}\rho_3(B)$ is the
1358: identity.
1359: Thus we have
1360: \[
1361: \overline{\sigma_3(t,B)}\sigma_3(t,B)=\id\;.
1362: \]
1363: By Lemma~\ref{fact}, the matrix $\sigma_3(t,B)$ can be written in the
1364: following form
1365: \begin{equation}\label{eq:form}
1366: \sigma_3(t,B)=
1367: \begin{pmatrix}
1368: p(t,B) & i \nu_1(t,B) \\
1369: i\nu_2(t,B) & \overline{p(t,B)}
1370: \end{pmatrix}
1371: \qquad \text{with }
1372: \nu_1,\nu_2\in \R \; , \quad
1373: p\bar p+\nu_1\nu_2=1 \; .
1374: \end{equation}
1375: We also have
1376: \begin{align*}
1377: F_{t,B}\circ \hat\mu_2 \circ \hat\mu_3
1378: &=\overline{\overline{F_{t,B}\circ \hat\mu_2 \circ \hat\mu_3}}
1379: =\overline{\sigma_3(t,B)\bigl(F_{t,B}\circ \hat\mu_2\bigr)\rho_3(B)^{-1}}\\
1380: &=
1381: \overline{\sigma_3(t,B)}\,\sigma_2(t,B)\,F_{t,B}\,\rho_2(B)^{-1}\,
1382: \overline{\rho_3(B)^{-1}}\;.
1383: \end{align*}
1384: We may assume that the small disk centered at $V_1$ consists of
1385: $2l$-copies of $D$.
1386: Let $b$ be the branching order of $g$ at $V_1$.
1387: By the condition \ref{item:gen:branch}, we have
1388: \begin{equation}\label{eq:ang-rel}
1389: A=\pi \frac{b+1}{l} \; .
1390: \end{equation}
1391: Since $\hat\mu_2 \circ \hat\mu_3$ is the rotation of angle $2A$ at
1392: $V_1$,
1393: we have $(\hat\mu_2 \circ \hat\mu_3)^l=\id$:
1394: \[
1395: F_{t,B}=F_{t,B}\circ (\hat\mu_2 \circ \hat\mu_3)^l
1396: =
1397: \left(\overline{\sigma_3(t,B)}\,\sigma_2(t,B)\right)^l
1398: F_{t,B}
1399: \left(\overline{\rho_3(B)}\,\rho_2(B)\right)^{-l}.
1400: \]
1401: On the other hand, one can easily check that the eigenvalues of
1402: $\overline{\rho_3(B)}\rho_2(B)$ are $\{-e^{iA},-e^{-iA}\}$.
1403: Then by \eqref{eq:ang-rel}, the eigenvalues of
1404: $\left(\overline{\rho_3(B)}\rho_2(B)\right)^{l}$ are $\{\pm 1,\pm 1\}$,
1405: that is
1406: \[
1407: \left(\overline{\rho_3(B)}\rho_2(B)\right)^{l}=\pm\id\;.
1408: \]
1409: So we have
1410: \[
1411: F_{t,B}=\pm \left(\overline{\sigma_3(t,B)}\sigma_2(t,B)\right)^l
1412: F_{t,B}
1413: \]
1414: which implies that
1415: \[
1416: \left(\overline{\sigma_3(t,B)}\sigma_2(t,B)\right)^l=\pm \id\;.
1417: \]
1418: This implies that the eigenvalues of
1419: $\overline{\sigma_3(t,B)}\sigma_2(t,B)$ are of the form
1420: $\{e^{\pi i N/l},e^{-\pi i N/l}\}$ for some integer $N$.
1421: Since $\overline{\sigma_3(t,B)}\sigma_2(t,B)$ is continuous with
1422: respect to the parameter $t$,
1423: we can conclude that the eigenvalues of
1424: $\overline{\sigma_3(t,B)}\sigma_2(t,B)$ do not change by $t$.
1425: In particular,
1426: \begin{equation}\label{eq:trace1}
1427: \trace\overline{\sigma_3(t,B)}\sigma_2(t,B)
1428: =\trace\overline{\sigma_3(0,B)}\sigma_2(0,B)\;.
1429: \end{equation}
1430: On the other hand, since
1431: $F_{t,B}\circ \hat\mu_2 \circ \hat\mu_3
1432: =\overline{\sigma_3(t,B)}\sigma_2(t,B)\,F_{t,B}\,
1433: (\overline{\rho_3(B)}\rho_2(B))^{-1}$ and $F_{0,B}=\id$,
1434: we have
1435: \begin{equation}\label{eq:trace2}
1436: \overline{\sigma_3(0,B)}\sigma_2(0,B)=
1437: \overline{\rho_3(B)}\rho_2(B)\;.
1438: \end{equation}
1439: By \eqref{eq:trace1}, \eqref{eq:trace2} and \eqref{eq:form}, we have
1440: \[
1441: 2 \Im p(t,B)=\trace\overline{\sigma_3(t,B)}\,\sigma_2(t,B)
1442: =\trace\overline{\rho_3(B)}\,\rho_2(B)
1443: =2 \cos A\;.
1444: \]
1445: \renewcommand{\qedsymbol}{}
1446: \end{proof}
1447: \begin{proof}[Step 4{\rm :}]
1448: Since $\sigma_3(0,B)=\rho_3(B)$, we have
1449: \[
1450: \Re p(0,B_0)=-\cos B_0\;.
1451: \]
1452: Now we would like to find a real valued smooth function $B(t)$ such
1453: that
1454: \[
1455: \Re p(t,B(t))=-\cos B_0 \qquad (B(0)=B_0)\;.
1456: \]
1457: By the implicit function theorem, a sufficient condition for
1458: the existence of such a $B(t)$ is
1459: \[
1460: \left.
1461: \frac{\partial \Re p(t,B)}{\partial B}
1462: \right |_{(t,B)=(0,B_0)} \ne 0\;,
1463: \]
1464: and by \eqref{eq:form}, \eqref{eq:q}, \eqref{eq:C} and the assumption
1465: \ref{item:gen:angle}, we have
1466: \begin{align*}
1467: \left.
1468: \frac{\partial \Re p(t,B)}{\partial B}
1469: \right |_{(t,B)=(0,B_0)}
1470: &=\left.
1471: \frac{\partial \Re p(0,B)}{\partial B}
1472: \right |_{B=B_0} \\
1473: &=\left.
1474: \frac{\partial \Re q(B)}{\partial B}
1475: \right |_{B=B_0} \\
1476: &=\left.
1477: -\frac{\partial \cos B}{\partial B}
1478: \right |_{B=B_0}=\sin B_0\ne 0\;.
1479: \end{align*}
1480: This proves the existence of such a $B(t)$
1481: $(|t|<\varepsilon)$ for a sufficiently
1482: small $\varepsilon>0$.
1483: \renewcommand{\qedsymbol}{}
1484: \end{proof}
1485: \begin{proof}[Step 5{\rm :}]
1486: When $t=0$, it holds that $\sigma_3(t,B)=\rho_3(B)$, so
1487: $\nu_1\nu_2>0$
1488: for sufficiently small $t$ ($|t|<\varepsilon$), by continuity.
1489: Now we set
1490: \[
1491: F_{t}:=
1492: \begin{pmatrix}
1493: u(t) & 0 \\
1494: 0 & u(t)^{-1}
1495: \end{pmatrix}
1496: F_{t,B(t)} \; ,
1497: \qquad
1498: u(t)=\sqrt[4]{\frac{\nu_2(t,B(t))}{\nu_1(t,B(t))}}
1499: \;.
1500: \]
1501: Obviously, $F_t$ satisfies the ordinary differential equation
1502: \eqref{eq:our_ode} for $B=B(t)$.
1503: In particular, $F_t$ has the Hopf differential $t\hat Q$ and the
1504: secondary Gauss map $\hat g_{B(t)}$.
1505: Then by \eqref{eq:A} and \eqref{eq:form}, we get
1506: the following relations
1507: \[
1508: \overline{F_t \circ \mu_j}
1509: =\sigma_j(t)\, F_t\, \rho_j(B(t)) \qquad (j=1,2,3)\;,
1510: \]
1511: where
1512: \begin{align*}
1513: \sigma_1(t)&=\id,\qquad
1514: \sigma_2(t)=\rho_2=
1515: \begin{pmatrix}
1516: -i & 0\\
1517: \hphantom{-}0 & i
1518: \end{pmatrix}\\
1519: \intertext{and}
1520: \sigma_3(t)&=
1521: \begin{pmatrix}
1522: p & i \sqrt{\nu_1\nu_2} \\
1523: i\sqrt{\nu_1\nu_2} & \bar p
1524: \end{pmatrix}\qquad
1525: \bigl(
1526: p = p(t,B(t)),~\nu_j=\nu_j(t,B(t)),~j=1,2
1527: \bigr)\;.
1528: \end{align*}
1529: Since $\Im p(t,B(t))=\cos A$ and $\Re p(t,B(t))=-\cos B_0$, we have
1530: \[
1531: p(t,B(t))=i\cos A-\cos B_0=q(B_0) \;.
1532: \]
1533: Thus we have $\sigma_j(t)=\rho_j(B_0)$ for $j=1,2,3$.
1534: We now set
1535: \[
1536: f_t:=F_tF_t^*\colon\widetilde M\longrightarrow H^3\;.
1537: \]
1538: By \eqref{eq:first}, the first fundamental form of $f_t$ is given by
1539: \[
1540: ds^2:=\left(1+|\hat g_{B(t)}|^2\right)^2
1541: \left|\frac{\hat Q}{d\hat g_{B(t)}}\right|^2\;.
1542: \]
1543: By the condition \ref{item:gen:branch} of the theorem, it is positive
1544: definite,
1545: and thus $f_t$ is a conformal \cmcone{} immersion whose
1546: Hopf differential is $t\hat Q$ and the secondary Gauss map
1547: $\hat g_{B(t)}$.
1548: \renewcommand{\qedsymbol}{}
1549: \end{proof}
1550: \begin{proof}[Step 6{\rm :}]
1551: We shall now prove that the conformal \cmcone{} immersion $f_t$
1552: is single-valued on $M=\overline M\setminus\{p_1,\dots,p_n\}$.
1553: Let $s$ be a non-negative integer and
1554: $\hat \mu_{i_1},\hat \mu_{i_2}, \dots , \hat \mu_{i_{2s}}$
1555: are sequences of three reflections $\hat \mu_1,\hat \mu_2,\hat \mu_3$
1556: on $M$ such that
1557: \[
1558: \pi\circ \hat \mu_{i_1}\circ \hat \mu_{i_2}\circ \cdots \circ
1559: \hat \mu_{i_{2s}}=\pi\;,
1560: \]
1561: where $\pi\colon{}\widetilde M\to M$ be the universal covering.
1562: To show the single-valued property of $f_t$, it is sufficient to show
1563: that
1564: $f_t=
1565: f_t\circ \hat\mu_{i_1}\circ
1566: \hat\mu_{i_2}\circ \cdots\circ \hat\mu_{i_{2s}}$.
1567: In fact, we have
1568: \begin{multline*}
1569: F_t\circ \hat\mu_{i_1}\circ
1570: \hat\mu_{i_2}\circ \dots \circ \hat\mu_{i_{2s}}
1571: = \left(\overline{\rho_{i_1}(B_0)}\rho_{i_2}(B_0)
1572: \dots \overline{\rho_{i_{2s-1}}(B_0)}\rho_{i_{2s}}(B_0)\right)
1573: F_t \\
1574: \left(\overline{\rho_{i_1}(B(t))}\rho_{i_2}(B(t))
1575: \dots \overline{\rho_{i_{2s-1}}(B(t))}\rho_{i_{2s}}(B(t))\right)\;.
1576: \end{multline*}
1577: Since $g$ is single-valued on $M$, $\hat g:=g\circ\pi$ satisfies
1578: \[
1579: \hat g=\hat g\circ \hat\mu_{i_1}\circ \cdots \circ \hat\mu_{i_{2s}}\;.
1580: \]
1581: On the other hand, since $\hat g$ is the secondary Gauss map of
1582: $F_{t,B_0}$,
1583: we have by \eqref{eq:three} that
1584: \[
1585: \hat g=
1586: \hat g\circ \hat\mu_{i_1}\circ \cdots \circ \hat\mu_{i_{2s}}
1587: = \left(\overline{\rho_{i_1}(B_0)}\,\rho_{i_2}(B_0)
1588: \dots \overline{\rho_{i_{2s-1}}(B_0)}\,\rho_{i_{2s}}(B_0)\right)\star g\;.
1589: \]
1590: Thus we can conclude that
1591: \[
1592: \overline{\rho_{i_1}(B_0)}\,\rho_{i_2}(B_0)
1593: \dots \overline{\rho_{i_{2s-1}}(B_0)}\,\rho_{i_{2s}}(B_0)
1594: =\pm \id\;.
1595: \]
1596: Since $\rho_{i_{j}}(B(t))\in \SU(2)$ ($j=1,\dots,2s$), this implies that
1597: $f_t=F_tF_t^*$ is single-valued on $M$.
1598:
1599: Moreover, by the assumption \ref{item:gen:ord} and
1600: Lemma~\ref{lem:irreg}, each end is irregular.
1601: The Hopf differential of $f_t$ is $Q_t$, and the pseudometric
1602: $d\sigma^2_{f_t}$ defined in \eqref{eq:g-metric} is
1603: $d\sigma^2_{A,B(t),\pi/2}$.
1604: Since they are symmetric with respect to $D$, \eqref{eq:gauss} and
1605: \eqref{eq:first} imply that the first and second fundamental forms
1606: of $f_t$ are invariant under the reflections $\mu_j$ $(j=1,2,3)$.
1607: Then by the fundamental theorem of surfaces, $f_t$ is symmetric with
1608: respect to $D$.
1609: \end{proof}
1610: \begin{proof}[Proof of Theorem~\ref{thm:irred}]
1611: Let $\tau$ be a loop surrounding the point $V_2$ with the base point
1612: $V_3$
1613: and $T$ the covering transformation of $\widetilde M$ corresponding
1614: to $\tau$.
1615: Suppose that a neighborhood of $V_2$ is generated by $2k$-copies of
1616: $D$.
1617: Then
1618: \[
1619: T := (\hat \mu_3\circ\hat \mu_1) ^ k
1620: \]
1621: holds.
1622: Thus we have
1623: \begin{align*}
1624: F_{t,B}\circ\tau &=
1625: F_{t,B}\circ(\hat\mu_3\circ\hat\mu_1)^k\\
1626: &=\left(\overline{\sigma_3(t,B)}\,\sigma_1\right)^k
1627: F_{t,B}
1628: \left(\rho_1^{-1}\overline{\rho_3(B)^{-1}}\right)^k
1629: = \overline{\sigma_3(t,B)}^k F_{t,B}\,\overline{\rho_3(B)}^{-k}\;.
1630: \end{align*}
1631: Here, by the argument of Step 6 of the proof of Theorem~\ref{thm:gen},
1632: we have
1633: $\rho_3(B)^k=\pm\id$.
1634: Hence at the base point $V_3$,
1635: \begin{align}\label{eq:tdiff}
1636: \left.\frac{\partial F_{t,B}\circ \tau}{\partial t}
1637: \right|_{(t,B)=(0,B_0)}
1638: &=
1639: \left.\frac{\partial}{\partial t}
1640: \overline{\sigma_3(t,B)}^k F_{t,B}\,\overline{\rho_3(B)}^{-k}
1641: \right|_{(t,B)=(0,B_0)}\\
1642: &=\pm\left.\frac{\partial}{\partial t}
1643: \overline{\sigma_3(t,B)}^k
1644: \right|_{(t,B)=(0,B_0)}
1645: \nonumber
1646: \end{align}
1647: since $F_{t,B}(V_3)=\id$.
1648:
1649: Since $F_{t,B}$ is a solution of
1650: \eqref{eq:our_ode}, it holds that
1651: \begin{align*}
1652: \frac{\partial}{\partial z}
1653: \left[\left.\frac{\partial F_{t,B}}{\partial t}
1654: \right|_{(t,B)=(0,B_0)}\right]\,dz
1655: &=
1656: d
1657: \left[\left.\frac{\partial F_{t,B}}{\partial t}
1658: \right|_{(t,B)=(0,B_0)}\right]\\
1659: &=
1660: \left.\frac{\partial }{\partial t} dF_{t,B}\right|_{(t,B)=(0,B_0)}\\
1661: &=
1662: \left.\frac{\partial}{\partial t}
1663: \left[
1664: t F_{t,B}
1665: \begin{pmatrix}
1666: {\hat g}_B & -{\hat g}_B^2 \\
1667: 1 & -{\hat g}_B
1668: \end{pmatrix}\frac{\hat Q}{d\hat g_B}
1669: \right]\right|_{(t,B)=(0,B_0)} \\
1670: &=
1671: \begin{pmatrix}
1672: \hat g & -\hat g^2 \\
1673: 1 & -\hat g\hphantom{^2}
1674: \end{pmatrix}\frac{\hat Q}{d\hat g} \; ,
1675: \end{align*}
1676: because $\hat g_{B_0}=\hat g=g\circ\pi$.
1677: Then by the assumption and \eqref{eq:tdiff},
1678: we have
1679: \[
1680: \left.\frac{\partial}{\partial t}
1681: \overline{\sigma_3(t,B)}^k
1682: \right|_{(t,B)=(0,B_0)}
1683: =\pm\oint_{\tau}
1684: \begin{pmatrix}
1685: g & -g^2 \\
1686: 1 & -g\hphantom{^2}
1687: \end{pmatrix}\frac{Q}{dg} \neq 0\;.
1688: \]
1689: This implies that $\sigma_3(t)=\sigma_3(t,B(t))$ is not constant
1690: on $\{t\,;\,|t|<\varepsilon\}$ for sufficiently small $\varepsilon>0$,
1691: and hence we have
1692: \begin{equation}\label{eq:c-irred}
1693: B(t)\neq B_0 \qquad \text{for}\quad 0<|t|<\varepsilon\;.
1694: \end{equation}
1695: Hence the eigenvalues of $\rho_3(t):=\rho_3(B(t))$ are different from
1696: those of $\rho_3(B_0)$.
1697:
1698: The secondary Gauss map $\hat g_t$ of $F_t$ changes by the covering
1699: transformation $T$ as
1700: \[
1701: \hat g_t\circ T = \hat g_t\circ(\hat\mu_3\circ\hat\mu_1)^k
1702: = \overline{\rho_3(t)}^k\star \hat g_t\;.
1703: \]
1704: Now, let $\widetilde V_2=\hat\mu_2(V_2)$
1705: (see Figure~\ref{fig:endreflect}).
1706: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
1707: \begin{figure}
1708: \setlength{\unitlength}{5mm}
1709: \small
1710: \begin{picture}(11,7)(-5,-1)
1711: \Thicklines
1712: \put(0,4){\line(1,-1){4}}
1713: \put(0,0){\line(1,0){4}}
1714: \put(0,0){\line(0,1){4}}
1715: \thinlines
1716: \put(-4,0){\line(1,1){4}}
1717: \put(-4,0){\line(1,0){4}}
1718: \put(0.2,-0.2){\makebox(0,0)[lt]{$V_3$}}
1719: \put(4,-0.2){\makebox(0,0)[lt]{$V_2$}}
1720: \put(-0.2,4){\makebox(0,0)[rb]{$V_1$}}
1721: \put(-4,-0.2){\makebox(0,0)[lt]{$\widetilde V_2$}}
1722: \put(2,0){\vector(0,1){0.5}}
1723: \put(2,0){\vector(0,-1){0.5}}
1724: \put(0,2){\vector(1,0){0.5}}
1725: \put(0,2){\vector(-1,0){0.5}}
1726: \put(2,2){\vector(1,1){0.35355339}}
1727: \put(2,2){\vector(-1,-1){0.35355339}}
1728: \put(-2,2){\vector(-1,1){0.35355339}}
1729: \put(-2,2){\vector(1,-1){0.35355339}}
1730: \put(2.2,-0.3){\makebox(0,0)[lt]{$\hat\mu_1$}}
1731: \put(-0.3,1.8){\makebox(0,0)[rt]{$\hat\mu_2$}}
1732: \put(2.6,2){\makebox(0,0)[lt]{$\hat\mu_3$}}
1733: \put(-3.2,2){\makebox(0,0)[lt]{$\tilde\mu_3$}}
1734: \end{picture}
1735: \caption{$\widetilde V_2$ and $\tilde\mu_3$.}
1736: \label{fig:endreflect}
1737: \end{figure}
1738: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
1739: We denote by $\tilde\mu_3$ the reflection about the edge $V_1\widetilde V_2$.
1740: Then we have
1741: \[
1742: \tilde\mu_3=\hat\mu_2\circ\hat\mu_3\circ\hat \mu_2\;.
1743: \]
1744: Let $\tilde \tau$ be a loop surrounding $\widetilde V_2$ with base
1745: point $V_3$, and let $\widetilde T$ be the covering transformation
1746: corresponding to $\tilde \tau$. Then we have
1747: \[
1748: \widetilde T = (\hat\mu_1\circ\tilde\mu_3)^k=
1749: (\hat\mu_1\circ\hat\mu_2\circ\hat\mu_3\circ\hat\mu_2)^k\;,
1750: \]
1751: and
1752: \[
1753: \hat g_t\circ\widetilde T =
1754: \left(\overline{\mathstrut\rho_1}\,\rho_2\,
1755: \overline{\rho_3(t)}\,\rho_2\right)^k
1756: \star \hat g_t
1757: = (-1)^{k-1}\rho_2\left(\overline{\rho_3(t)}^k\right)
1758: \rho_2\star \hat g_t\;.
1759: \]
1760: So, to show irreducibility, it is sufficient to show that the
1761: matrices
1762: \[
1763: a:=\overline{\rho_3(t)}^k
1764: \qquad \text{and}\qquad
1765: b:=\rho_2\overline{\rho_3(t)}^k\rho_2 = \rho_2\,a\,\rho_2
1766: \]
1767: do not commute.
1768: By \eqref{eq:rho}, $\overline{\rho_3(t)^k}\rho_3(t)^k=\id$ holds.
1769: Then by Lemma~\ref{fact}, the off-diagonal components of $a$ are
1770: coincide and pure imaginary .
1771: Set
1772: \[
1773: a=\overline{\rho_3(t)}^k = \begin{pmatrix}
1774: r & i\beta \\
1775: i\beta & \bar r
1776: \end{pmatrix}
1777: \qquad (\beta\in\R,\,r\bar r+\beta^2=1)\;.
1778: \]
1779: Then we have
1780: \[
1781: b = \rho_2 \, a \,\rho_2 =
1782: \begin{pmatrix}
1783: -r & i\beta \\
1784: i\beta & -\bar r
1785: \end{pmatrix}\;,
1786: \]
1787: and
1788: \[
1789: \left[a,b\right] = a b - b a
1790: =\begin{pmatrix}
1791: 0 & -2\beta\Im r\\
1792: 2\beta\Im r & 0
1793: \end{pmatrix}\;,
1794: \]
1795: that is, $a$ and $b$ commute if and only if $\beta=0$ or $r$ is a
1796: real number.
1797:
1798: First, we consider the case $\beta=0$.
1799: Then $a=\overline{\rho_3(t)}^k$ is a diagonal matrix whose eigenvalues
1800: are different from $\pm 1$ for sufficiently small $t$, because $B(t)$ is
1801: not constant.
1802: In particular, the two eigenvalues of $a$ are distinct.
1803: This implies that the eigenspaces of $a$ coincide of those of
1804: $\overline{\rho_3(t)}$.
1805: Since $a$ is diagonal, this implies that $\rho_3(t)$ is also a diagonal
1806: matrix, a contradiction.
1807:
1808: Next, assume $r$ is real.
1809: Then there exists a real number $\theta$ such that
1810: \begin{multline*}
1811: a = \overline{\rho_3(t)}^k
1812: =\begin{pmatrix}
1813: \cos\theta & i\sin\theta \\
1814: i\sin\theta & \cos\theta
1815: \end{pmatrix}
1816: =P\begin{pmatrix} e^{i\theta} & 0 \\ 0 & e^{-i\theta}\end{pmatrix}
1817: P^{-1}\;,\\
1818: \text{where }
1819: P=\frac{1}{\sqrt{2}}
1820: \begin{pmatrix}
1821: 1 & -1 \\
1822: 1 & \hphantom{-}1
1823: \end{pmatrix}\; \qquad \text{and}\qquad
1824: \theta\in\R\setminus\pi\Z\;.
1825: \end{multline*}
1826: In this case, $(P^{-1}\overline{\rho_3(t)}P)^k$ is a diagonal matrix
1827: whose eigenvalues differ from $\pm 1$, for sufficiently small $t\neq 0$.
1828: Then, by a similar argument to the previous case, we have
1829: $P^{-1}\overline{\rho_3}(t)P$ is diagonal.
1830: If $A\not\equiv\pi/2 \pmod{\pi}$, this contradicts to \eqref{eq:rho}
1831: and \eqref{eq:C}.
1832: Hence $a$ and $b$ do not commute.
1833: \end{proof}
1834: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
1835: \section{An example of genus one}
1836: \label{sec:torus}
1837: In the final section, we construct an example of a \cmcone{} surface
1838: of genus one with four irregular ends.
1839:
1840: Let $\Gamma=\Z\oplus i\Z$ be the lattice of Gaussian integers of
1841: $\C$ and let
1842: \[
1843: \overline M:= \C/2\Gamma\;.
1844: \]
1845: We consider $\overline M$ as the square
1846: $[-\frac{1}{2},\frac{3}{2}]\times[-\frac{1}{2},\frac{3}{2}]$
1847: in $\C=\R^2$, with opposite edges identified.
1848: Take a triangle $D$ on $\overline M$ as in Figure~\ref{fig:fund-torus}.
1849: Then $\overline M$ is obtained from $D$ by reflecting across the
1850: edges of $D$.
1851: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
1852: \begin{figure}
1853: \setlength{\unitlength}{30mm}
1854: \small
1855: \begin{center}
1856: \begin{picture}(3,3)(-1,-1)
1857: \put(-1,0){\vector(1,0){2.8}}
1858: \put(0,-1){\vector(0,1){2.8}}
1859: \drawline(-0.5,-0.5)(-0.5,1.5)(1.5,1.5)(1.5,-0.5)(-0.5,-0.5)
1860: \drawline(-0.5,1)(1.5,1)
1861: \drawline(1,-0.5)(1,1.5)
1862: \drawline(-0.5,0.5)(1.5,0.5)
1863: \drawline(0.5,-0.5)(0.5,1.5)
1864: \Thicklines
1865: \drawline(0,0)(0.5,0)(0.5,0.5)(0,0)
1866: \put(0,0){\circle*{0.05}}
1867: \put(1,0){\circle*{0.05}}
1868: \put(0,1){\circle*{0.05}}
1869: \put(1,1){\circle*{0.05}}
1870: \put(0.3,0.15){\makebox(0,0){$D$}}
1871: \put(0.55,-0.1){\makebox(0,0)[lc]{$V_3=\frac{1}{2}$}}
1872: \put(0.55,0.55){\makebox(0,0)[lb]{$V_1=\frac{i}{2}$}}
1873: \put(-0.05,-0.1){\makebox(0,0)[rc]{$V_2=0$}}
1874: \put(0.1,-0.1){\makebox(0,0)[lc]{$p_1$}}
1875: \put(1.1,-0.1){\makebox(0,0)[lc]{$p_2$}}
1876: \put(1.1,1.1){\makebox(0,0)[lc]{$p_3$}}
1877: \put(0.1,1.1){\makebox(0,0)[lc]{$p_4$}}
1878: \end{picture}
1879: \end{center}
1880: \caption{Fundamental domain of the torus}
1881: \label{fig:fund-torus}
1882: \end{figure}
1883: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
1884:
1885: Using the Weierstrass $\wp$ function with respect to $\Gamma$
1886: (not with respect to $2\Gamma$),
1887: we set
1888: \[
1889: Q = \bigl(\wp'(z)\bigr)^2\,dz^2\;.
1890: \]
1891: Then $Q$ has poles at $\{p_1,p_2,p_3,p_4\}=\{0,1,1+i,i\}$, each
1892: with order $6$.
1893: The $\wp$-function with respect to the square lattice has the following
1894: properties
1895: \[
1896: \overline{\wp(\bar z)}=\wp(z)\;,\qquad
1897: \overline{\wp(-\bar z)}=\wp(z)\;,\qquad
1898: \overline{\wp(i\bar z)}=-\wp(z)\;.
1899: \]
1900: Hence $Q$ is symmetric with respect to $D$.
1901:
1902: Consider an abstract spherical triangle $\Trig(A,B_0,C)$
1903: with
1904: \[
1905: A = \frac{3}{4}\pi\;,\qquad
1906: B_0 = \frac{\pi}{2}\;,\qquad
1907: C = \frac{3}{2}\pi\;,
1908: \]
1909: and identify the triangle with the fundamental region $D$.
1910: Then the metric of $\Trig(A,B_0,C)$ can be extended to
1911: $d\sigma^2_{A,B_0,C}\in\metone(\overline M)$ by reflections.
1912: Since $A$, $B_0$ and $C$ satisfy \eqref{eq:tri}, $d\sigma^2_{A,B_0,C}$
1913: is non-degenerate.
1914: Let $g$ be the developing map of $d\sigma^2_{A,B_0,C}$.
1915: Since the conical orders of $d\sigma^2_{A,B_0,C}$ are integers, $g$ is
1916: well-defined on $\C$.
1917:
1918: Now, we prove that $g$ is well-defined on $\overline M$.
1919: By the monodromy principle (Lemma \ref{lem:p-I}), one can choose
1920: $g$ such that $\overline{g\circ\hat\mu_j}=\rho_j\star
1921: g$ ($j=1,2,3$), where
1922: \[
1923: \rho_1:=\id,\qquad
1924: \rho_2:=\begin{pmatrix}
1925: i & \hphantom{-}0 \\
1926: 0 & -i
1927: \end{pmatrix},\qquad
1928: \rho_3:=\frac{1}{\sqrt{2}}
1929: \begin{pmatrix}
1930: \hphantom{-}i & \pm i \\
1931: \pm i & -i
1932: \end{pmatrix}
1933: \]
1934: and $\hat\mu_1$, $\hat\mu_2$ and $\hat\mu_3$ are reflections along
1935: the edges $V_3V_1$, $V_1V_2$ and $V_2V_3$, respectively.
1936: We denote by $\tau_1$ and $\tau_2$ the translations $z\mapsto z+1$
1937: and $z\mapsto z+i$ respectively.
1938: Then
1939: \[
1940: \tau_1=\hat\mu_2\circ\hat\mu_3\circ\hat\mu_1\circ\hat\mu_3,
1941: \qquad
1942: \tau_2=\hat\mu_3\circ\hat\mu_2\circ\hat\mu_3\circ\hat\mu_1
1943: \]
1944: holds.
1945: So we have
1946: \begin{align*}
1947: g\circ\tau_1 &= \overline{\rho_2}\,\rho_3\,\overline{\rho_1}\,\rho_3
1948: \star g
1949: = \rho_2\star g\;,\\
1950: g\circ\tau_2 &= \overline{\rho_3}\,\rho_2\,\overline{\rho_3}\,\rho_1
1951: \star g
1952: = \pm\begin{pmatrix}
1953: 0 & i \\ i & 0
1954: \end{pmatrix}\star g\;.
1955: \end{align*}
1956: Thus
1957: \[
1958: g(z+2)=g\circ\tau_1\circ\tau_1(z) = g(z)\;,\qquad
1959: g(z+2i)=g\circ\tau_2\circ\tau_2(z) =g(z)
1960: \]
1961: hold.
1962: This shows that $g$ is invariant under the action of the lattice
1963: $2\Gamma$.
1964: Hence $g$ is a meromorphic function on $\overline M$.
1965:
1966: One can easily see that the same result as Theorem~\ref{thm:gen}
1967: holds when $C=3\pi/2$, instead of $\pi/2$. Hence
1968: we have a one-parameter family $\{f_t\}$ of \cmcone{} immersions of
1969: $\overline M\setminus\{p_1,p_2,p_3,p_4\}$ into $H^3$ with irregular
1970: ends.
1971:
1972:
1973: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
1974: \appendix
1975: \begingroup
1976: \renewcommand{\thesection}{\Alph{section}}
1977: \section{}
1978: \label{app:A}
1979: \endgroup
1980: For a compact Riemann surface $\overline M$ and points
1981: $p_1,\dots,p_n\in \overline M$, a conformal metric $d\sigma^2$ of
1982: constant curvature $1$ on $M:=\overline M\setminus\{p_1,\dots,p_n\}$ is
1983: an element of $\metone(\overline M)$ if there exist real numbers
1984: $\beta_1,\dots,\beta_n >-1$ so that each $p_j$ is a conical singularity
1985: of conical order $\beta_j$, that is, if $d\sigma^2$ is
1986: asymptotic to $c_j|z-p_j|^{2\beta_j}\,dz\cdot d\bar z$
1987: at $p_j$, for $c_j\neq 0$ and $z$ a local complex coordinate around
1988: $p_j$.
1989: We call the formal sum
1990: \begin{equation}\label{eq:divisor}
1991: D:=\sum_{j=1}^n \beta_j\, p_j
1992: \end{equation}
1993: the {\em divisor\/} corresponding to $d\sigma^2$.
1994: For a pseudometric $d\sigma^2\in\metone(\overline M)$ with divisor $D$,
1995: there is a holomorphic map $g\colon{}\widetilde M\to\CP^1$ such that
1996: $d\sigma^2$ is the pull-back of the Fubini-Study metric of $\CP^1$.
1997: This map, called the {\em developing map\/} of $d\sigma^2$, is uniquely
1998: determined up to M\"obius transformations $g \mapsto a\star g$ for
1999: $a \in \SU(2)$.
2000: We have the following expression
2001: \[
2002: \pi^*d\sigma^2=\frac{4\,dg\cdot d\bar g}{(1+|g|^2)^2}\;,
2003: \]
2004: where $\pi:\widetilde M\to M$ is a covering projection.
2005:
2006: Consider $d\sigma^2\in\metone(\overline M)$ with divisor $D$ as in
2007: \eqref{eq:divisor} and with the developing map $g$.
2008: Since the pull-back of the
2009: Fubini-Study metric of $\CP^1$ by $g$ is invariant under the deck
2010: transformation group $\pi_1(M)$ of
2011: $M:=\overline M\setminus\{p_1,\dots,p_n\}$, there is a representation
2012: \[
2013: \rho_g\colon{}\pi_1(M)\longrightarrow
2014: \SU(2)
2015: \]
2016: such that
2017: \[
2018: g\circ T^{-1} = \rho_g(T)\star g \qquad (T\in\pi_1(M))\;.
2019: \]
2020: The metric $d\sigma^2$ is called {\em reducible\/} if the image
2021: of $\rho_g$ is a commutative subgroup in $\SU(2)$, and is called
2022: {\em irreducible\/} otherwise.
2023: Since the maximal abelian subgroup of $\SU(2)$ is $\U(1)$,
2024: the image of $\rho_g$ for a reducible $d\sigma^2$ lies in a subgroup
2025: conjugate to $\U(1)$, and this image might be simply the identity.
2026: We call a reducible metric $d\sigma^2$ {\em $\Hyp^3$-reducible\/}
2027: if the image of $\rho_g$ is the identity, and {\em $\Hyp^1$-reducible\/}
2028: otherwise (for more on this, see \cite[Section~3]{ruy1}).
2029:
2030: The following assertion was needed in Section~\ref{sec:reflection}:
2031: \begin{proposition}\label{prop:A}
2032: Let $d\sigma^2$ be a metric of constant curvature $1$ defined on
2033: $M=\overline{M}
2034: \setminus \{p_1,\dots,p_n\}$ whose conical order at each $p_j$ is an
2035: integer.
2036: Then the developing map $g$ of $d\sigma^2$ is single-valued
2037: on the universal covering of $M$.
2038: \end{proposition}
2039: Let $p_1,\dots,p_{n-1}$ be distinct points in $\C$ and $p_n=\infty$.
2040: We set
2041: \[
2042: M_{p_1,\dots,p_n}:=\C\cup\{\infty\}
2043: \setminus \{p_1,p_2,\dots,p_n\} \qquad (p_n=\infty)
2044: \]
2045: and $\widetilde{M}_{p_1,\dots,p_n}$ its universal covering.
2046: \begin{corollary}\label{cor:A}
2047: Let $d\sigma^2$ be a metric of constant curvature $1$ defined on
2048: $M_{p_1,\dots,p_n}$ $(p_n=\infty)$ whose conical order at each
2049: $p_j$ is an integer.
2050: Then the developing map $g$ of $d\sigma^2$ is single-valued on
2051: $M_{p_1,\dots,p_n}$ and extends as a meromorphic function on
2052: $\C\cup\{\infty\}$.
2053: Moreover, the divisor of $d\sigma^2$ coincides with the ramification
2054: divisor of the meromorphic function $g$.
2055: \end{corollary}
2056:
2057: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
2058: \begin{thebibliography}{RUY5}
2059: \bibitem[BB]{BB}
2060: J. L. Barbosa and P. Berard,
2061: {\itshape Eigenvalue and
2062: ``twisted'' eigenvalue problems, applications to CMC surfaces},
2063: J. Math.\ Pures Appl. {\bfseries 79} (2000), 427--450.
2064: \bibitem[Bry]{Bryant}
2065: R.~Bryant,
2066: {\itshape Surfaces of mean curvature one in hyperbolic space},
2067: Ast\'erisque {\bfseries 154--155} (1987), 321--347.
2068: \bibitem[BR]{BR}
2069: J. Berglund and W. Rossman,
2070: {\itshape Minimal surfaces with catenoid ends},
2071: Pacific J. Math. {\bfseries 171} (1995), 353--371.
2072: \bibitem[CGT]{CGT}
2073: M. P. do Carmo, J. de M. Gomes and G. Thorbergsson,
2074: {\itshape The influence of the boundary behaviour on hypersurfaces with
2075: constant mean curvature in $H^{n+1}$},
2076: Comment. Math. Helvetici {\bfseries 61} (1986), 429--441.
2077: \bibitem[CL]{CL}
2078: M. P. do Carmo and H. B. Lawson,
2079: {\itshape On Alexandrov-Berstein theorems in hyperbolic space},
2080: Duke Math. J. {\bfseries 50(4)} (1983), 995--1003.
2081: \bibitem[CS]{CS}
2082: M. P. do Carmo and A. M. da Silveira,
2083: {\itshape Index and total curvature of surfaces with constant mean
2084: curvature},
2085: Proc. Amer.\ Math.\ Soc. {\bfseries 110(4)} (1990), 1009--1015.
2086: %\bibitem[CG]{cg}
2087: % C.~C.~Chen, F.~Gackstatter,
2088: % {\itshape Elliptische und hyperelliptische Funktionen und
2089: % vollst\"andige Minimalfl\"achen vom Enneperschen Typ},
2090: % Math.~Ann. {\bfseries 259} (1982), 359--369.
2091: \bibitem[CHR1]{chr}
2092: P.~Collin, L.~Hauswirth and H.~Rosenberg,
2093: {\itshape The geometry of finite topology Bryant surfaces},
2094: Ann.\ of Math. {\bfseries 153} (2001), 623--659.
2095: \bibitem[CHR2]{chr2}
2096: \bysame,
2097: {\itshape The gaussian image of mean curvature one surfaces in
2098: $H^3$ of finite total curvature},
2099: Advanced Studies in Pure Mathematics
2100: 29, Minimal Surfaces, Geometric Analysis and Symplectic Geometry 2000.
2101: \bibitem[CN]{CN}
2102: C. J. Costa and V. F. de Sousa Neto,
2103: {\itshape Mean curvature $1$ surfaces of Costa type in hyperbolic
2104: 3-space},
2105: T\^ohoku Math.\ J.\ {\bfseries 53} (2001), 617--628.
2106: \bibitem[D]{D}
2107: B. Daniel,
2108: {\itshape Surfaces de Bryant dans $H^3$ de type fini},
2109: Bull.\ Sci.\ Math.\ {\bfseries 126} (2002), 581--594.
2110: \bibitem[ET1]{ET1}
2111: R. Sa Earp and E. Toubiana,
2112: {\itshape On the geometry of constant mean curvature one surfaces in
2113: hyperbolic space},
2114: Illinois J. Math. {\bfseries 45} (2001), 371--401.
2115: \bibitem[ET2]{ET2}
2116: \bysame,
2117: {\itshape Meromorphic data for mean curvature one surfaces in
2118: hyperbolic space},
2119: Tohoku Math. J. (2) 56 (2004), no. 1, 27--64.
2120: \bibitem[F]{F}
2121: M. Furuta,
2122: {\itshape Gluing spherical triangles}
2123: (Ky\^umen sankaku kei wo hariawaseru to),
2124: in Japanese,
2125: S\^ugaku Ts\^ushin {\bf 6} (2001), 4--27.
2126: \bibitem[FH]{FH}
2127: M. Furuta and Y. Hattori,
2128: {\itshape Two-dimensional spherical space forms}, preprint.
2129: \bibitem[GGN]{ggn}
2130: C. C. G\'oes, M. E. E. L. Galv\~ao, B. Nelli,
2131: {\itshape A type Weierstrass representation for surfaces in
2132: hyperbolic space with mean curvature equal to one},
2133: An. Acad. Brasil Cienc. 70 (1998), no. 1, 1-6.
2134: \bibitem[Kar]{Kar}
2135: H. Karcher,
2136: {\itshape Hyperbolic constant mean curvature one
2137: surfaces with compact fundamental domains},
2138: Global theory of minimal surfaces, 311--323, Clay Math. Proc. 2, Amer.
2139: Math. Soc., 2005.
2140: \bibitem[Kat]{Kat}
2141: S. Kato,
2142: {\itshape Construction of $n$-end catenoids with prescribed flux},
2143: Kodai Math.~J {\bfseries 18} (1995), 86--98.
2144: \bibitem[KUY]{kuy}
2145: M. Kokubu, M. Umehara and K. Yamada,
2146: {\itshape An elementary proof of Small's formula for null curves
2147: in $\operatorname{PSL}(2,\C)$
2148: and an analogue for Legendrian curves in
2149: $\operatorname{PSL}(2,\C)$},
2150: Osaka J. Math. 40 (2003), no. 3, 697--715
2151: \bibitem[LR]{LR}
2152: L. L. de Lima and W. Rossman,
2153: {\itshape On the index of mean
2154: curvature $1$ surfaces in $H^3$},
2155: Indiana Univ.\ Math.\ J. {\bfseries 47(2)} (1998), 685--723.
2156: \bibitem[MU]{MU}
2157: C. McCune and M. Umehara,
2158: {\itshape An analogue of the UP-iteration for constant mean curvature
2159: one surfaces in hyperbolic 3-space},
2160: Diff. Geom. Appl. 20 (2004), no. 2, 197--207.
2161: \bibitem[PP]{PP}
2162: F. Pacard and F. A. Pimentel,
2163: {\itshape Attaching handles to Bryant surfaces},
2164: J. Inst. Math. Jussieu 3 (2004), no. 3, 421--459.
2165: \bibitem[RR]{RR}
2166: L. Rodriguez and H. Rosenberg,
2167: {\itshape Half-space theorems for mean curvature one surfaces
2168: in hyperbolic space},
2169: Proc.~Amer.~Math.~Soc. {\bfseries 126} (1998), 2763--2771.
2170: \bibitem[Ros]{Ro}
2171: H. Rosenberg,
2172: {\itshape Bryant surfaces},
2173: The global theory of minimal surfaces in flat spaces
2174: (Martina Franca, 1999),
2175: pp. 67--111, Lecture Notes in Math., Vol.\ 1775, Springer-Verlag,
2176: 2002.
2177: \bibitem[RS]{rs}
2178: W.~Rossman and K.~Sato,
2179: {\itshape Constant mean curvature surfaces with two ends in hyperbolic
2180: space},
2181: Experimental Math. {\bfseries 7(2)} (1998), 101--119.
2182: \bibitem[RUY1]{ruy1}
2183: W.~Rossman, M.~Umehara and K.~Yamada,
2184: {\itshape Irreducible constant mean curvature 1 surfaces in
2185: hyperbolic space with positive genus},
2186: T\^ohoku Math.~J. {\bfseries 49} (1997), 449--484.
2187: \bibitem[RUY2]{ruy2}
2188: \bysame,
2189: {\itshape A new flux for mean curvature $1$ surfaces
2190: in hyperbolic $3$-space, and applications},
2191: Proc.~Amer.~Math.~Soc. {\bfseries 127} (1999), 2147--2154.
2192: \bibitem[RUY3]{ruy3}
2193: \bysame,
2194: {\itshape Mean curvature $1$ surfaces with low total curvature
2195: in hyperbolic $3$-space I}, Hiroshima Math. J. 34 (2004), 21--56.
2196: \bibitem[RUY4]{ruy4}
2197: \bysame,
2198: {\itshape Mean curvature $1$ surfaces with low total curvature
2199: in hyperbolic $3$-space II}, T\^ohoku J.\ Math. 55 (2003), 375--395.
2200: \bibitem[RUY5]{ruy5}
2201: \bysame,
2202: {\itshape Period problems for mean curvature one surfaces in $H^3$
2203: {\rm (}with application to surfaces of low total curvature{\rm})},
2204: Adv. Stud. Pure Math. 51 (2008), 335--387.
2205: \bibitem[Sm]{Sm}
2206: A.~J.~Small,
2207: {\itshape Surfaces of constant mean curvature $1$ in $H^3$ and
2208: algebraic curves on a quadric},
2209: Proc.~Amer.~Math.~Soc.
2210: {\bfseries 122} (1994), 1211--1220.
2211: \bibitem[Tro1]{Troyanov1}
2212: M.~Troyanov,
2213: {\itshape Metric of constant curvature on a sphere with two conical
2214: singularities},
2215: in ``Differential Geometry'', Lect. Notes in Math. vol.~1410,
2216: Springer-Verlag, (1989), 296--306.
2217: \bibitem[UY1]{uy1}
2218: M.~Umehara and K.~Yamada,
2219: {\itshape Complete surfaces of constant mean curvature-$1$
2220: in the hyperbolic $3$-space},
2221: {Ann. of Math. {\bfseries 137} (1993), 611--638.}
2222: \bibitem[UY2]{uy2}
2223: \bysame,
2224: {\itshape A parameterization of Weierstrass formulae and
2225: perturbation of some complete minimal surfaces of
2226: $\R^3$ into the hyperbolic $3$-space},
2227: J. reine u.~angew.~Math. {\bfseries 432} (1992), 93--116.
2228: \bibitem[UY3]{uy3}
2229: \bysame,
2230: {\itshape Surfaces of constant mean curvature-$c$
2231: in $H^3(-c^2)$ with prescribed hyperbolic Gauss map},
2232: Math. Ann. {\bfseries 304} (1996), 203--224.
2233: \bibitem[UY4]{uy4}
2234: \bysame,
2235: {\itshape Another construction of a CMC $1$ surface in $H^3$},
2236: Kyungpook Math. J. {\bfseries 35} (1996), 831--849.
2237: \bibitem[UY5]{uy5}
2238: \bysame,
2239: {\itshape A duality on CMC $1$ surfaces in hyperbolic $3$-space
2240: and a hyperbolic analogue of the Osserman Inequality},
2241: Tsukuba J. Math. {\bfseries 21} (1997), 229-237.
2242: \bibitem[UY6]{uy6}
2243: \bysame,
2244: {\itshape Geometry of surfaces of constant mean curvature $1$ in the
2245: hyperbolic $3$-space},
2246: Suugaku Expositions {\bfseries 10(1)} (1997), 41--55.
2247: \bibitem[UY7]{uy7}
2248: \bysame,
2249: {\itshape Metrics of constant curvature $1$ with three conical
2250: singularities on the $2$-sphere},
2251: Illinois J. Math. {\bfseries 44(1)} (2000), 72--94.
2252: \bibitem[Xu]{Xu}
2253: Y. Xu,
2254: {\itshape Symmetric minimal surface in $\R^3$},
2255: Pacific J. Math. {\bfseries 171} (1995), 203--224.
2256: \bibitem[Yu1]{Yu1}
2257: Z.~Yu,
2258: {\itshape Value distribution of hyperbolic Gauss maps},
2259: Proc.~Amer.~Math.~Soc.
2260: {\bfseries 125} (1997), 2997--3001.
2261: \bibitem[Yu2]{Yu2}
2262: \bysame,
2263: {\itshape The inverse surface and the Osserman Inequality},
2264: Tsukuba J. Math. {\bfseries 22} (1998), 575--588
2265: \bibitem[Yu3]{Yu3}
2266: \bysame,
2267: {\itshape Surfaces of constant mean curvature one in the hyperbolic
2268: three-space with irregular ends},
2269: T\^ohoku Math.\ J. {\bfseries 53} (2001) 305--318.
2270: \end{thebibliography}
2271: \end{document}
2272:
2273: