1: \documentclass[a4paper]{article}
2: \usepackage{amsmath,amsthm,amssymb}
3: \usepackage[dvips]{graphicx}
4:
5:
6: \headheight=0mm
7: \headsep=0mm
8: \topmargin=0mm
9: \oddsidemargin=0mm
10: \textwidth=160mm
11: \textheight=230mm
12:
13: \newtheorem{theorem}{Theorem}
14: \newtheorem{lemma}[theorem]{Lemma}
15: \newtheorem{proposition}[theorem]{Proposition}
16: \newtheorem{remark}[theorem]{Remark}
17:
18: \begin{document}
19: \title{Special Solutions of the Sixth Painlev\'e Equation \\
20: with Solvable Monodromy}
21: \author{Kazuo Kaneko \\
22: \small Graduate School of Information Science and Technology, %\\
23: \small Osaka University\\[+12pt]
24: Shoji Okumura\\
25: \small Graduate School of Science, Osaka University
26: }
27: \date{}
28: \maketitle
29: \begin{abstract}
30: We will study two types of special solutions of the sixth Painlev\'e
31: equation, which are invariant under the symmetries obtained from the
32: B\"acklund transformations. In most cases, the fixed points of the
33: B\"acklund transformations are classical solution, but our solutions
34: are not classical for generic parameters. We will calculate the linear
35: monodromy of these solutions exactly, and we will characterize them on
36: Fricke's cubic surface of monodromy.
37: \end{abstract}
38: \section{Introduction}
39: R. Fuchs showed that the sixth Painlev\'e equation is represented as an
40: isomonodromic deformation condition of the Fuchs type linear equation
41: with four regular singular points~\cite{RF1}. Garnier showed that every
42: type of the Painlev\'e equation is also obtained from monodromy preserving
43: deformation of linear equation~\cite{Gar}.
44: Nevertheless it is difficult to calculate the monodromy itself for the
45: generic Painlev\'e functions.
46:
47: We call the monodromy data of the linear equation as a \textit{linear
48: monodromy} of the Painlev\'e function.
49: Our interests is in the Painlev\'e function whose corresponding linear
50: monodromy can be determined exactly. In this paper, we call such
51: Painlev\'e functions \textit{monodromy solvable}.
52:
53: One example of monodromy solvable Painlev\'e functions is Umemura's
54: classical solutions~\cite{HU1}, which are fixed points of the B\"acklund
55: transformations. For any classical solutions, we can calculate the
56: linear monodromy.
57: But conversely, there exist some non-classical Painlev\'e functions
58: whose linear monodromy can be calculated.
59: We are interested in non-classical monodromy solvable solutions.
60:
61: R. Fuchs is the first to find a non-classical monodromy solvable
62: solution~\cite{RF2}.
63: He calculates the linear monodromy of Picard's solution \cite{EP},
64: which satisfies the sixth Painlev\'e equation with a special parameter.
65: His work is found again in \cite{MMA} recently.
66:
67: Another example of monodromy solvable solution is a symmetric solution.
68: The first, second and fourth Painlev\'e equations
69: \begin{alignat*}{2}
70: &(P1)&\qquad &\frac{d^2 y}{dt^2}=6y^2+t,\\
71: &(P2)&\qquad &\frac{d^2 y}{dt^2}=2y^{3}+ty+\alpha,\\
72: &(P4)&\qquad &\frac{d^2
73: y}{dt^2}=\frac{1}{2y}\left(\frac{dy}{dt}\right)^{2}+\frac{3}{2}
74: y^3+4ty^{2}+2(t^{2}-\alpha)y+\frac\beta y
75: \end{alignat*}
76: are invariant under the transformations
77: \begin{alignat*}{4}
78: &(P1)&\qquad &t\to \rho t,&\quad y&\to \rho^3 y& \quad(\rho^5=1), \\
79: &(P2)&\qquad &t\to \omega t,&\quad y&\to \omega^2 y&\quad
80: (\omega^3=1), \\
81: &(P4)&\qquad &t \to -t,&\quad y&\to -y.
82: \end{alignat*}
83: There exist symmetric solutions which are invariant under the action of
84: the cyclic groups. The symmetric solutions are studied by
85: Kitaev~\cite{AVK} for (P1) and (P2), and by Kaneko~\cite{KK1} for (P4).
86: For these cases, $(y,t)=(0,0)$ is a fixed point of the transformations,
87: and there exist symmetric solutions with initial values at
88: the fixed point $y(0)=0$. Kitaev calculated the linear monodromy of the
89: symmetric solutions of (P1) and (P2) by adopting the fixed point as the
90: initial condition. In the same way, Kaneko calculated the linear
91: monodromy of the symmetric solution of (P4). We remark that these
92: symmetric solutions are not classical for generic parameters.
93:
94: In this paper, we will construct monodromy solvable solutions of the
95: sixth Painlev\'e equation.
96:
97: For generic parameter, we cannot find a simple symmetry as above.
98: However, for some restricted parameters, there exist two types of
99: symmetries obtained from the B\"acklund transformations.
100: One is
101: \[
102: \sigma_{1}:\ t\to 1-t,\qquad y\to 1-y,\qquad \beta \to -\gamma,\qquad \gamma \to -\beta,
103: \]
104: and the other is
105: \[
106: \sigma_{2}\circ\sigma_{1}:\ t\to 1/(1-t),\qquad y\to 1/(1-y),\qquad
107: \alpha\to-\beta,\qquad
108: \beta\to-\gamma,\qquad
109: \gamma\to\alpha.
110: \]
111: The sixth Painlev\'e equation is invariant under the action $\sigma_{1}$
112: if $-\beta=\gamma$, and is invariant under the action
113: $\sigma_{2}\circ\sigma_{1}$ if $\alpha=-\beta=\gamma$.
114:
115: In the section 3, we will show there exist symmetric solutions, which
116: are fixed points of these actions.
117: In many cases, the fixed points of the B\"acklund transformations
118: are classical solutions. However, our symmetric solutions are not
119: classical for generic parameters.
120:
121: In the section 4, we will show that the linearizations of these symmetric
122: solutions are reduced to the Gauss hypergeometric equations. In the
123: section 5, we will calculate the linear monodromy of our symmetric solutions
124: explicitly.
125:
126: If the monodromy matrices $\left\{M_{0},M_{t},M_{1},M_{\infty}\right\}$
127: are elements of $SL(2,\mathbb{C})$, then
128: $$
129: p_{j}=\mathrm{tr}M_{j}, \qquad p_{jk}=\mathrm{tr} M_{j}M_{k},
130: $$
131: satisfies the following relation:
132: \begin{multline}
133: p_{01}p_{1t}p_{t0}+p_{01}^2+p_{1t}^2+p_{t0}^2
134: -(p_{0}p_{1}+p_{t}p_{\infty})
135: p_{01}-(p_{1}p_{t}+p_{0}p_{\infty})p_{1t}-(p_{t}p_{0}+p_{1}p_{\infty})p_{t0}
136: \\
137: +p_{0}^2+p_{1}^2+p_{t}^2+p_{\infty}^{2}
138: +p_{0}p_{t}p_{1}p_{\infty}-4=0.\label{eq-mon}
139: \end{multline}
140: It is known that the monodromy matrices
141: $\{M_{0},M_{t},M_{1},M_{\infty}\}$ are
142: determined by
143: $\{p_{0}, p_{1}, p_{t}, p_{\infty}; p_{01}, p_{1t}, p_{t0}\}$
144: upto gauge transformations(\cite{Boa}, \cite{Iwa}, \cite{Jim}, \cite{Fri}).
145: %Therefore the monodromy is on the cubic surface \eqref{eq-mon}.
146: We call \eqref{eq-mon} as Fricke's cubic surface of monodromy.
147: In the section 6, we will
148: characterize the monodromy of our symmetric solutions on Fricke's cubic
149: surface of monodromy.
150:
151: Andreev and Kitaev constructed many algebraic solutions of
152: the sixth Painlev\'e equation, by rational transformations of the
153: hypergeometric equations~\cite{AK}. By the rational transformations of
154: confluent hypergeometric equation, Ohyama and Okumura constructed
155: algebraic solutions and symmetric solutions of the
156: first to the fifth Painlev\'e equations~\cite{Oh}.
157: We remark that Boach also constructed many algebraic solutions of the
158: sixth Painlev\'e equation, whose linear monodromy is isomorphic to the
159: complex reflection group~\cite{Boa}.
160:
161: The solutions of this paper essentially appeared
162: in~\cite{AK}. But, since their interest are in algebraic solutions, they
163: did not go to detail of these solutions. In this paper, we will
164: calculate the linear monodromy explicitly, and we will show that they
165: are symmetric solutions.
166:
167: \bigskip
168:
169: The authors give thanks to Professor Yousuke Ohyama
170: for fruitful discussions and suggestions.
171:
172: \section{Linear Problem}\label{LP}
173: The sixth Painlev\'e equation \eqref{p6} is obtained from the
174: isomonodromic deformation equation of
175: linear equation with four regular singular points. We will use
176: Garnier-Okamoto's linearization \cite{KO}.
177: \begin{align}
178: &\frac{\partial^2 \psi}{\partial x^2}+p(x,t)\frac{\partial \psi}{\partial
179: x}+q(x,t)\psi=0, \label{OG-1}\\
180: &\frac{\partial\psi}{\partial t}=a(x,t)\frac{\partial \psi}{\partial x}+
181: b(x,t)\psi, \label{OG-2}
182: \end{align}
183: where
184: \begin{align}
185: p(x,t)&=\frac{1-\alpha_{4}}{x}+\frac{1-\alpha_{3}}{x-1}+\frac{1-\alpha_{0}}{x-t}-\frac{1}{x-y},\\
186: q(x,t)&=\frac{\alpha_{2}(\alpha_{1}+\alpha_{2})}{x(x-1)}-
187: \frac{t(t-1)H_{VI}}{x(x-1)(x-t)}+\frac{y(y-1)z}{x(x-1)(x-y)},\\
188: a(x,t)&=\frac{y-t}{t(t-1)}\frac{x(x-1)}{x-y},\\
189: b(x,t)&=\frac{(1-\alpha_{4}-\alpha_{3}-\alpha_{0})(y-t)}{2t(t-1)}-
190: \frac{y(y-1)(y-t)z}{t(t-1)(x-y)},
191: \end{align}
192: and
193: \begin{multline}
194: H_{VI}=\frac 1{t(t-1)}\Bigl[
195: y(y-1)(y-t)z^2-\Bigl\{\alpha_4(y-1)(y-t)\\{}+\alpha_3y(y-t)+(\alpha_0-1)y(y-1)
196: \Bigr\}z
197: +\alpha_2(\alpha_{1}+\alpha_{2})(y-t)
198: \Bigr].
199: \end{multline}
200: The Riemann scheme of \eqref{OG-1} is
201: \[\left\{x;
202: \begin{matrix}
203: 0 & 1 & t & y & \infty \\
204: 0 & 0 & 0 & 0 & \alpha_{2}\\
205: \alpha_{4}&\alpha_{3}&\alpha_{0}&2&\alpha_{1}+\alpha_{2}
206: \end{matrix}\right\}.
207: \]
208: From the compatibility condition of \eqref{OG-1} and \eqref{OG-2},
209: we have the following Hamiltonian system
210: \[
211: \frac{dy}{dt}=\frac{\partial H_{VI}}{\partial z}, \qquad
212: \frac{dz}{dt}=-\frac{\partial H_{VI}}{\partial y},
213: \]
214: that is
215: \begin{align}
216: \begin{aligned}
217: t(t-1)\frac{dy}{dt}=&2y(y-1)(y-t)z-
218: (\alpha_{0}-1)y(y-1)-
219: \alpha_3y(y-t)-
220: \alpha_4(y-1)(y-t),\\[+3pt]
221: t(t-1)\frac{dz}{dt}=&-\Bigl(y(y-1)+(y-1)(y-t)+y(y-t)\Bigr)z^2 \\
222: &\qquad\qquad
223: +\Bigl((2y-1)(\alpha_{0}-1)+(2y-t)\alpha_{3}+(2y-t-1)\alpha_{4}\Bigr)z-
224: (\alpha_1 + \alpha_2)\alpha_2 .
225: \end{aligned}\label{P6H}
226: \end{align}
227: Let us eliminate $z$ from this system, and we have the sixth Painlev\'e
228: equation
229: \begin{multline}
230: \frac{d^2 y}{d t^2} = \frac{1}{2}\left( \frac{1}{y} +
231: \frac{1}{y-1} + \frac{1}{y-t} \right) \left(\frac{d y}{d t}\right)^2
232: - \left( \frac{1}{t}+\frac{1}{t-1}
233: +\frac{1}{y-t} \right)\frac{d y}{d t} \\
234: + \frac{y\left(y-1\right)\left(y-t\right)}{t^2\left(t-1\right)^2}
235: \left\{ \alpha + \beta\frac{t}{y^2} +
236: \gamma\frac{t-1}{\left(y-1\right)^2 }+
237: \delta \frac{t\left(t-1\right)}{\left(y-t\right)^2} \right\},
238: \label{p6}
239: \end{multline}
240: where
241: \[
242: \alpha=\frac{\alpha_{1}^{2}}{2},\qquad
243: \beta=-\frac{\alpha_{4}^{2}}{2},\qquad
244: \gamma=\frac{\alpha_{3}^{2}}{2},\qquad
245: \delta=\frac{1-\alpha_{0}^{2}}{2}.
246: \]
247:
248: \bigskip
249:
250: We will calculate the linear monodromy of some special Painlev\'e
251: functions.
252:
253: \section{B\"acklund Transformations of the Sixth Painlev\'e Equation}
254: \label{sym}
255: For the sixth Painlev\'e equation,
256: we cannot find a simple symmetry for general parameters. However, for
257: some restricted parameters, there exist some symmetries obtained from
258: the B\"acklund transformations. And there exist symmetric solutions
259: of the B\"acklund transformations.
260:
261: \bigskip
262:
263: The sixth Painlev\'e equation \eqref{P6H} is invariant under the
264: following transformations\cite{NY}:
265: \begin{center}
266: \begin{tabular}{|c||ccccc|cc|c|}\hline
267: {} &$ \alpha_0 $&$ \alpha_1 $&$ \alpha_2 $&$ \alpha_3 $&$ \alpha_4 $&$ y $&$ z $&$ t$\\
268: \hline
269: $s_0 $&$ -\alpha_0 $&$ \alpha_1 $&$ \alpha_0+\alpha_2 $&$ \alpha_3 $&$ \alpha_4 $&$ y $&$ z-\frac{\alpha_0}{y-t} $&$ t$\\$
270: s_1 $&$ \alpha_0 $&$ -\alpha_1 $&$ \alpha_1+\alpha_2 $&$ \alpha_3 $&$ \alpha_4 $&$ y $&$ z$&$ t$\\$s_2 $&$ \alpha_0+\alpha_2 $&$ \alpha_1+\alpha_2 $&$ -\alpha_2
271: $&$ \alpha_2+\alpha_3 $&$ \alpha_2+\alpha_4 $&$ y+\frac{\alpha_2}{z} $&$ z $&$ t$\\$
272: s_3 $&$ \alpha_0 $&$ \alpha_1 $&$ \alpha_2+\alpha_3 $&$ -\alpha_3 $&$ \alpha_4 $&$ y
273: $&$ z-\frac{\alpha_3}{y-1} $&$ t$\\$s_4 $&$ \alpha_0 $&$ \alpha_1 $&$ \alpha_2+\alpha_4
274: $&$ \alpha_3 $&$ -\alpha_4 $&$ y $&$ z-\frac{\alpha_4}{y} $&$ t$\\
275: \hline
276: $ \pi_1 $&$ \alpha_3$&$ \alpha_4 $&$ \alpha_2 $&$ \alpha_0 $&$ \alpha_1 $&$ \frac{t}{y}
277: $&$ -\frac{y\left(yz+\alpha_2\right)}{t} $&$ t$\\
278: $\pi_2 $&$ \alpha_1 $&$ \alpha_0 $&$ \alpha_2 $&$ \alpha_4 $&$ \alpha_3
279: $&$ \frac{t\left(y-1\right)}{y-t}
280: $&$
281: -\frac{\left(y-t\right)^2z+\left(y-t\right)\alpha_2}{\left(t-1\right)
282: t}$&$ t$\\
283: \hline
284: $\sigma_1 $&$ \alpha_0 $&$ \alpha_1 $&$ \alpha_2 $&$ \alpha_4 $&$ \alpha_3 $&$ 1-y $&$ -z $&$ 1-t$\\$
285: \sigma_2 $&$ \alpha_0 $&$ \alpha_4 $&$ \alpha_2 $&$ \alpha_3 $&$ \alpha_1 $&$ \frac{1}{y} $&$ -y\left(yz+\alpha_2\right) $&$ \frac{1}{t}$\\$
286: \sigma_3 $&$ \alpha_4 $&$ \alpha_1 $&$ \alpha_2 $&$ \alpha_3 $&$ \alpha_0 $&$ \frac{t-y}{t-1} $&$ \left(1-t\right)z $&$ \frac{t}{t-1}$\\
287: \hline
288: \end{tabular}\end{center}
289: where $\alpha_{0},\alpha_{1},\alpha_{2},\alpha_{3}$ and $\alpha_{4}$ are the
290: parameters in \eqref{P6H}, and satisfy the following relation:
291: \[
292: \alpha_{0}+\alpha_{1}+2\alpha_{2}+\alpha_{3}+\alpha_{4}=1.
293: \]
294: We remark that
295: $$\langle s_{0}, s_{1}, s_{2}, s_{3}, s_{4}, \pi_{1}, \pi_{2}\rangle$$
296: is an extended affine Weyl group of type $D_{4}^{(1)}$, and
297: $$\langle s_{0}, s_{1}, s_{2}, s_{3}, s_{4}, \pi_{1}, \pi_{2},
298: \sigma_{1},\sigma_{2},\sigma_{3}\rangle$$ is an extended affine Weyl
299: group of type $F_{4}^{(1)}$.
300:
301: \bigskip
302:
303: If $\alpha_{3}=\alpha_{4}$, \eqref{P6H} is invariant under
304: the action of $\sigma_{1}$, and if $\alpha_{1}=\alpha_{3}=\alpha_{4}$,
305: \eqref{P6H} is invariant under the action of $\sigma_{2}\circ\sigma_{1}$.
306: In the following, we will study the symmetric solutions of $\sigma_{1}$
307: and $\sigma_{2}\circ\sigma_{1}$.
308:
309:
310: \subsection{Symmetric Solution of $\sigma_{1}$}
311:
312: For the B\"acklund transformation $\sigma_{1}$:
313: \[
314: y \to 1-y, \qquad z \to -z, \qquad t \to 1-t,
315: \]
316: the sixth Painlev\'e equation is invariant with the condition
317: $\alpha_{3}=\alpha_{4}$, and one of the fixed point is
318: \[
319: t=1/2, \quad y=1/2, \quad z=0.
320: \]
321: Since this is not a critical point of $\eqref{P6H}$,
322: there exist a unique holomorphic solution of \eqref{P6H} with the
323: initial condition
324: \[
325: y(1/2)=1/2, \qquad z(1/2)=0.
326: \]
327: And if $\alpha_{3}=\alpha_{4}$, this solution is expanded as follows:
328: \begin{align}
329: y(t)=\sum_{k=0}^\infty a_{k}\left(t-\tfrac 12\right)^{k},\qquad
330: z(t)=\sum_{k=0}^{\infty} b_{k}\left(t-\tfrac 12\right)^{k},
331: \label{sym2}
332: \end{align}
333: where
334: \begin{align*}
335: &a_{0}=\frac 12,\quad a_{1}=1-\alpha_{0},\quad a_{2}=0,\\
336: &\qquad a_{3}=-\frac 43 \alpha_{0}\bigl\{
337: 2\alpha_{2}(\alpha_{1}+\alpha_{2})-(1-\alpha_{0})(2\alpha_{2}+\alpha_{1}+1)
338: \bigr\},\quad a_{4}=0, \quad \cdots,\\[+2ex]
339: &b_{0}=0,\quad b_{1}=4\alpha_{2}(\alpha_{1}+\alpha_{2}),\quad b_{2}=0,\\
340: &\qquad
341: b_{3}=-\frac{16}3
342: \alpha_{2}(\alpha_{1}+\alpha_{2})\bigl\{
343: \alpha_{2}(\alpha_{1}+\alpha_{2})-(\alpha_{1}+2\alpha_{2})(1-2\alpha_{0})
344: +\alpha_{0}-2 \bigr\},\quad b_{4}=0, \quad
345: \cdots.
346: \end{align*}
347:
348: \begin{proposition}
349: The Painlev\'e function \eqref{sym2} is a symmetric solution of
350: $\sigma_{1}$:
351: \[
352: y(1-t)=1-y(t),\qquad
353: z(1-t)=-z(t).
354: \]
355: \end{proposition}
356: \begin{proof}
357: Let us set
358: \[
359: \tau=t-1/2, \qquad \lambda=y-1/2, \qquad \mu=z,
360: \]
361: then the action of $\sigma_{1}$ maps $\tau\to -\tau$, $\lambda\to
362: -\lambda$, and $\mu\to -\mu$.
363: \eqref{P6H} becomes the following Hamiltonian system
364: $\frac{d\lambda}{d\tau}=\frac{\partial
365: K}{\partial \mu}$, $\frac{d\mu}{d\tau}=-\frac{\partial K}{\partial
366: \lambda}$:
367: \begin{align}
368: \begin{aligned}
369: \left(\tau^2-\tfrac{1}{4}\right)\frac{d \lambda}{d \tau}&=
370: 2\left(\lambda^2-\tfrac{1}{4}\right)\mu\left(\lambda-\tau\right)
371: -\left(\lambda^2-\tfrac{1}{4}\right)\left(\alpha_{0}-1\right)
372: -2\lambda\left(\lambda-\tau\right)\alpha_{3},\\
373: \left(\tau^2-\tfrac{1}{4}\right)\frac{d \mu}{d \tau}&=
374: -\left(\left(\lambda^2-\tfrac{1}{4}\right)
375: +2\lambda\left(\lambda-\tau\right)\right)\mu^2
376: +2\left(\lambda(\alpha_{0}-1)+(2\lambda-\tau)\alpha_{3}
377: \right)\mu
378: -\alpha_{2}(\alpha_{1}+\alpha_{2}).
379: \end{aligned}\label{H2mod}
380: \end{align}
381: The Hamiltonian $K$ is
382: \[
383: \left(\tau^2-\tfrac 1 4\right)K=
384: \mu^2(\lambda-\tau)\left(\lambda^2-\tfrac 14\right)-\mu\left\{
385: \left(\lambda^2-\tfrac 1 4\right) (\alpha_{0}-1)+2\lambda(\lambda-\tau)\alpha_{3}\right\}+(\lambda-
386: \tau)\alpha_{2}(\alpha_{1}+\alpha_{2}) .
387: \]
388: Then \eqref{sym2} corresponds to the following solution of \eqref{H2mod}:
389: \begin{align*}
390: \lambda = (1-\alpha_0)\tau + O(\tau)^3,\qquad
391: \mu =4\alpha_2(\alpha_1+\alpha_2)\tau + O(\tau)^3.
392: \end{align*}
393: Higher order expansions are determined from
394: \eqref{H2mod} inductively, and it has only odd powers of $\tau$.
395: Therefore it determines symmetric solution. Thus \eqref{sym2} is a
396: symmetric solution.
397: \end{proof}
398:
399:
400: By the B\"acklund transformations $s_{0}$, $s_{1}$, $s_{2}$ and
401: $\pi_{2}$, which are commutative with $\sigma_{1}$, we have four
402: symmetric solutions around $t=1/2$.
403:
404: \begin{proposition} \label{sym2-prop}
405: If $\alpha_3=\alpha_4$,
406: all of the symmetric solutions on $\sigma_{1}$ around $t=1/2$ are
407: \begin{align}
408: &
409: \left\{
410: \begin{aligned}
411: y- \tfrac 12 &= (1-\alpha_0)\left(t-\tfrac 12\right)
412: + O\left(t-\tfrac 12\right)^3,
413: \\
414: z&=4\alpha_2(\alpha_1+\alpha_2)\left(t-\tfrac 12\right)
415: + O\left(t-\tfrac 12\right)^3,
416: \end{aligned}\right.\tag{S2-1}\label{sym2-1}
417: \\
418: &
419: \left\{
420: \begin{aligned}
421: y-\tfrac 12&= (1+\alpha_0)\left(t-\tfrac 12\right)
422: + O\left(t-\tfrac 12\right)^3,
423: \\
424: z&=\left(t-\tfrac 12\right)^{-1}
425: + O\left(t-\tfrac 12\right),
426: \end{aligned}\right.\tag{S2-2}\label{sym2-2}
427: \\
428: &\left\{
429: \begin{aligned}
430: y-\tfrac 12&= \tfrac 1{4\alpha_1}\left(t-\tfrac 12\right)^{-1}
431: +O\left(t-\tfrac 12\right),
432: \\
433: z&=-4\alpha_1\alpha_2\left(t-\tfrac 12\right)
434: + O\left(t-\tfrac 12\right)^3,
435: \end{aligned}\right.\tag{S2-3}
436: \label{sym2-3}
437: \\
438: &\left\{
439: \begin{aligned}
440: y-\tfrac 12&= \tfrac {-1}{4\alpha_1}\left(t-\tfrac 12\right)^{-1}
441: + O\left(t-\tfrac 12\right) ,
442: \\
443: z&=4\alpha_1(\alpha_1+\alpha_2)\left(t-\tfrac 12\right)
444: + O\left(t-\tfrac 12\right)^3 ,
445: \end{aligned}\right.\tag{S2-4}
446: \label{sym2-4}
447: \end{align}
448: where \eqref{sym2-1} is \eqref{sym2}.
449:
450:
451: The actions of $s_{0}$, $s_{1}$, $s_{2}$ and $\pi_{2}$ interchange the
452: symmetric solutions as follows:
453: \begin{center}
454: \begin{picture}(130,85)
455: \put(0,20){\makebox(10,10){$\scriptstyle s_1$}}
456: \put(15,25){\oval(8,8)[l]}
457: \put(15,21){\vector(1,0){4}}
458: \put(15,29){\vector(1,0){4}}
459: \put(20,20){\makebox(30,10){\eqref{sym2-2}}}
460: \put(50,20){\makebox(30,10){$\scriptstyle s_2$}}
461: \put(55,25){\oval(8,8)[r]}
462: \put(55,21){\vector(-1,0){4}}
463: \put(55,29){\vector(-1,0){4}}
464: \put(80,20){\makebox(30,10){\eqref{sym2-4}}}
465: \put(75,25){\oval(8,8)[l]}
466: \put(75,21){\vector(1,0){4}}
467: \put(75,29){\vector(1,0){4}}
468: \put(115,25){\oval(8,8)[r]}
469: \put(115,21){\vector(-1,0){4}}
470: \put(115,29){\vector(-1,0){4}}
471: \put(110,20){\makebox(30,10){$\scriptstyle s_0$}}
472: \put(35,10){\vector(0,1){9}}
473: \put(95,10){\vector(0,1){9}}
474: \put(35,10){\line(1,0){60}}
475: \put(60,0){\makebox(10,10){$\scriptstyle \pi_{2}$}}
476: \put(35,36){\vector(0,1){28}}
477: \put(35,64){\vector(0,-1){28}}
478: \put(36,45){\makebox(10,10){$\scriptstyle s_{0}$}}
479: \put(95,35){\vector(0,1){30}}
480: \put(95,65){\vector(0,-1){30}}
481: \put(96,45){\makebox(10,10){$\scriptstyle s_{1}$}}
482: \put(20,65){\makebox(30,10){\eqref{sym2-1}}}
483: \put(80,65){\makebox(30,10){\eqref{sym2-3}}}
484: \put(50,72){\vector(1,0){30}}
485: \put(50,68){\vector(1,0){30}}
486: \put(80,72){\vector(-1,0){30}}
487: \put(80,68){\vector(-1,0){30}}
488: \put(60,74){\makebox(10,10)[b]{$\scriptstyle s_{2}$}}
489: \put(60,55){\makebox(10,10)[t]{$\scriptstyle \pi_{2}$}}
490: \put(0,65){\makebox(10,10){$\scriptstyle s_1$}}
491: \put(15,70){\oval(8,8)[l]}
492: \put(15,66){\vector(1,0){4}}
493: \put(15,74){\vector(1,0){4}}
494: \put(115,70){\oval(8,8)[r]}
495: \put(115,66){\vector(-1,0){4}}
496: \put(115,74){\vector(-1,0){4}}
497: \put(110,65){\makebox(30,10){$\scriptstyle s_0$}}
498: \end{picture}
499: \end{center}
500: \end{proposition}
501:
502:
503:
504:
505:
506:
507:
508: In section \ref{MS2}, we will show these symmetric solutions
509: are monodromy solvable.
510:
511: \subsection{Symmetric Solution of $\sigma_{2}\circ \sigma_{1}$}
512: For the B\"acklund transformation $\sigma_{2}\circ\sigma_{1}$:
513: \[
514: y\to\frac 1{1-y},\qquad
515: z\to-(1-y)(-z(1-y)+\alpha_{2}),\qquad
516: t\to\frac 1{1-t},
517: \]
518: the sixth Painlev\'e equation is invariant with the condition
519: $\alpha_{1}=\alpha_{3}=\alpha_{4}$, and one of the fixed point is
520: \[
521: y=-\omega^2,\qquad
522: z=\frac{2\omega+1}3\alpha_{2},\qquad
523: t=-\omega^2.
524: \]
525: Since this is not a critical point of $\eqref{P6H}$,
526: there exist a unique holomorphic solution of \eqref{P6H} with the initial
527: condition
528: \[
529: y(-\omega^2)=-\omega^{2}, \qquad z(-\omega^2)=\frac{2\omega+1}3\alpha_{2}.
530: \]
531: And if $\alpha_{1}=\alpha_{3}=\alpha_{4}$, this solution is expanded as
532: follows:
533: \begin{align}
534: \begin{aligned}
535: &y(t)=-\omega^{2}+(1-\alpha_{0})(t+\omega^2)+\frac{1+2\omega}3\alpha_{0}(1-\alpha_{0})(t+\omega^2)^2+O(t+\omega^2)^3,\\
536: &z(t)=\frac{2\omega+1}3 \alpha_{2}-\frac{\alpha_{2}}3(1-\alpha_{0})
537: (t+\omega^2)+O(t+\omega^2)^2.
538: \end{aligned}\label{sym3}
539: \end{align}
540:
541: \begin{proposition}
542: The Painlev\'e function \eqref{sym3} is a symmetric solution of
543: $\sigma_{2}\circ\sigma_{1}$:
544: \[
545: y\left(\frac 1{1-t}\right)=\frac 1{1-y(t)},\qquad
546: z\left(\frac 1{1-t}\right)=-(1-y(t))\left\{-z(t)(1-y(t))+\alpha_{2}\right\}.
547: \]
548: \end{proposition}
549: \begin{proof}
550: To prove the symmetry of $\sigma_{2}\circ\sigma_{1}$, we set
551: \begin{align*}
552: \tau=\frac{-\omega t-1}{t+\omega},\qquad
553: \lambda=\frac{-\omega y-1}{y+\omega},
554: \end{align*}
555: which diagonalize the action of
556: $\sigma_2 \circ \sigma_1$: $\tau\to\omega\tau$,
557: $\lambda\to\omega\lambda$.
558: Then the sixth Painlev\'e equation \eqref{p6} becomes
559: \begin{multline}
560: (\lambda-\tau)\frac{d^2\lambda}{d\tau^2}=\frac{4\lambda^3-3\tau\lambda^2+1}
561: {2(\lambda^3+1)}\left(\frac{d\lambda}{d\tau}\right)^2+
562: \frac{-3\tau^2\lambda+2\tau^3-1}{\tau^3+1}
563: \frac{d\lambda}{d\tau}\\
564: {}+\frac{(\lambda^3+1)^2(\tau^3+1)(1-\alpha_0^2)+9(\lambda-\tau)^2
565: (\lambda^4-2\tau\lambda^3-2\lambda+\tau)\alpha_1^2}
566: {2(\lambda^3+1)(\tau^3+1)^2}, \label{P6 modified}
567: \end{multline}
568: which is equivalent to the following Hamiltonian system
569: $\frac{d\lambda}{d\tau}=\frac{\partial K}{\partial \mu}$,
570: $\frac{d\mu}{d\tau}=-\frac{\partial K}{\partial \lambda}$:
571: \begin{align}
572: \begin{aligned}
573: (\tau^{3}+1)\frac{d\lambda}{d\tau}=&
574: 2\left( 1 + \lambda^3 \right) \mu \left( \lambda - \tau \right)
575: +3\lambda^2\left( \lambda - \tau \right)\alpha_1
576: +\left( 1 + \lambda^3 \right) \left( 1 + \alpha_0 \right)
577: ,\\
578: (\tau^{3}+1)\frac{d\mu}{d\tau}=&
579: \mu^2\left( - 4\lambda^3 + 3\lambda^2\tau -1 \right)
580: -3\lambda \mu \left( \lambda \left( 1 + \alpha_0 \right) +
581: \left( 3\lambda - 2\tau \right) \alpha_1 \right)
582: - \lambda {\left( \alpha_2 -1\right) }^2 \\[-6pt]
583: &\hspace{5cm}+\frac{ \alpha_2-1 }{2}\left( \left( \lambda + \tau \right) \left( 1 + \alpha_0 \right) +
584: 3\left( \lambda - \tau \right) \alpha_1 \right) .
585: \end{aligned}\label{H6 modified}
586: \end{align}
587: The Hamiltonian $K$ is
588: \begin{multline*}
589: (\tau^{3}+1)K=\mu^2(\lambda^{3}+1)(\lambda-\tau)+
590: \mu\Bigl\{
591: (\lambda^3+1)(1+\alpha_{0})+3\lambda^2(\lambda-\tau)\alpha_{1}
592: \Bigr\}\\
593: +\frac{1+\alpha_{0}+3\alpha_{1}}4
594: \lambda\Bigl\{
595: (\lambda+\tau)(1+\alpha_{0})+3(\lambda-\tau)\alpha_{1}
596: \Bigr\},
597: \end{multline*}
598: and $\mu$ is determined by
599: \begin{multline*}
600: \mu=\frac{t(t-1)(y+\omega)^2}{2y(y-1)(y-t)(1-\omega^2)}\frac{dy}{dt}
601: +\frac{(\omega-1)(t+\omega)^2(\alpha_{0}+1)}{6(y-t)}+
602: \frac{(1-\omega^2)\alpha_{1}}{6y}+
603: \frac{\omega^2(1-\omega^2)\alpha_{1}}{6(y-1)}\\
604: +
605: \frac{(1-\omega)\alpha_{1}y}{6}
606: +\frac 1 6 \left\{
607: (\omega-1)(t+\omega)(1+\alpha_{0})+(\omega^2-1)\alpha_{1}
608: \right\}.
609: \end{multline*}
610:
611: By the action of $\sigma_{2}\circ\sigma_{1}$, we have
612: \[
613: \tau\to\omega\tau, \qquad
614: \lambda\to\omega\lambda,\qquad
615: \mu\to\omega^2\mu,\qquad
616: K\to\omega^2 K.
617: \]
618: There exists a solution of \eqref{H6 modified}:
619: \begin{align}
620: \begin{aligned}
621: \lambda&=\left( 1 - \alpha_0 \right) \tau %+
622: +O(\tau)^4,
623: \\
624: \mu&=\tau^{-1}+O(\tau)^2,
625: \end{aligned}\label{sym3mod}
626: \end{align}
627: which are corresponding to \eqref{sym3}. And inductively, we have
628: \eqref{sym3mod} in the following forms
629: \begin{align*}
630: &\lambda=\sum a_{n} \tau^{3n+1},
631: &\mu=\sum b_{n} \tau^{3n-1}.
632: \end{align*}
633: Therefore it is a symmetric solution. And thus \eqref{sym3} is a symmetric
634: solution.
635: \end{proof}
636:
637: By the B\"acklund transformations $s_{0}$ and $s_{2}$, which are
638: commutative with $\sigma_{2}\circ\sigma_{1}$, we have three symmetric
639: solutions around $t=-\omega^2$.
640:
641: \begin{proposition}
642: If $\alpha_{1}=\alpha_{3}=\alpha_{4}$,
643: all of symmetric solutions on $\sigma_{2}\circ\sigma_{1}$ around
644: $t=-\omega^2$ are
645: \begin{align}
646: &\left\{
647: \begin{aligned}
648: y&=-\omega+O(t+\omega^2)^{2},\\
649: z&=-\tfrac 1 3 (1+2\omega)\alpha_{2}
650: -\tfrac 1 6 (\alpha_{0}-3\alpha_{1}-1)\alpha_{2}(t+\omega^2)
651: +O(t+\omega^2)^{2},
652: \end{aligned}\right.\tag{S3-1}\label{sym3-1}\\
653: &\left\{
654: \begin{aligned}
655: y&=-\omega^2+(1-\alpha_{0})(t+\omega^2)+O(t+\omega^2)^2,\\
656: z&=\tfrac 1 3 (1+2\omega)\alpha_{2}+
657: \tfrac 1 3(\alpha_{0}-1)\alpha_{2}(t+\omega^2)+O(t+\omega^2)^{2},
658: \end{aligned}\right.\tag{S3-2}\label{sym3-2}\\
659: &\left\{
660: \begin{aligned}
661: y&=-\omega^2+(1+\alpha_{0})(t+\omega^2)+O(t+\omega^2)^{2},\\
662: z&=(t+\omega^2)^{-1}+\tfrac 1 2 (1+2\omega)(\alpha_{0}-\alpha_{1}+1)
663: +O(t+\omega^2),
664: \end{aligned}\right.\tag{S3-3}\label{sym3-3}
665: \end{align}
666: where \eqref{sym3-2} is \eqref{sym3}.
667: The actions of $s_{0}$ and $s_{2}$ interchange the symmetric solutions
668: as follows:
669: \begin{center}
670: \begin{picture}(190,15)
671: \put(0,0){\makebox(10,10){$\scriptstyle s_0$}}
672: \put(15,5){\oval(8,8)[l]}
673: \put(15,1){\vector(1,0){4}}
674: \put(15,9){\vector(1,0){4}}
675: \put(20,0){\makebox(30,10){\eqref{sym3-1}}}
676: \put(50,5){\makebox(30,10){$\scriptstyle s_2$}}
677: \put(50,5){\vector(1,0){30}}
678: \put(80,5){\vector(-1,0){30}}
679: \put(80,0){\makebox(30,10){\eqref{sym3-2}}}
680: \put(110,5){\makebox(30,10){$\scriptstyle s_0$}}
681: \put(110,5){\vector(1,0){30}}
682: \put(140,5){\vector(-1,0){30}}
683: \put(140,0){\makebox(30,10){\eqref{sym3-3}}}
684: \put(175,5){\oval(8,8)[r]}
685: \put(175,0){\vector(-1,0){4}}
686: \put(175,9){\vector(-1,0){4}}
687: \put(180,0){\makebox(10,10){$\scriptstyle s_2$}}
688: \end{picture}
689: \end{center}
690: There are more three symmetric solutions around $t=-\omega$:
691: \begin{align}
692: &\left\{
693: \begin{aligned}
694: y&=-\omega^2+O(t+\omega)^{2},\\
695: z&=-\tfrac 1 3 (1+2\omega^2)\alpha_{2}
696: -\tfrac 1 6 (\alpha_{0}-3\alpha_{1}-1)\alpha_{2}(t+\omega)
697: +O(t+\omega)^{2},
698: \end{aligned}\right.\tag{S3-4}\label{sym3-1p}
699: \\
700: &\left\{
701: \begin{aligned}
702: y&=-\omega+(1-\alpha_{0})(t+\omega)+O(t+\omega)^2,\\
703: z&=\tfrac 1 3 (1+2\omega^2)\alpha_{2}+
704: \tfrac 1 3(\alpha_{0}-1)\alpha_{2}(t+\omega)+O(t+\omega)^{2},
705: \end{aligned}\right.\tag{S3-5}\label{sym3-2p}
706: \\
707: &\left\{
708: \begin{aligned}
709: y&=-\omega+(1+\alpha_{0})(t+\omega)+O(t+\omega)^{2},\\
710: z&=(t+\omega)^{-1}+\tfrac 1 2 (1+2\omega^2)(\alpha_{0}-\alpha_{1}+1)
711: +O(t+\omega),
712: \end{aligned}\right.\tag{S3-6}\label{sym3-3p}
713: \end{align}
714: which correspond to \eqref{sym3-1}, \eqref{sym3-2} and \eqref{sym3-3}
715: by $\omega\to\omega^2$.
716: The actions of $s_{0}$ and $s_{2}$ interchange the symmetric solutions
717: as follows:
718: \begin{center}
719: \begin{picture}(190,15)
720: \put(0,0){\makebox(10,10){$\scriptstyle s_0$}}
721: \put(15,5){\oval(8,8)[l]}
722: \put(15,1){\vector(1,0){4}}
723: \put(15,9){\vector(1,0){4}}
724: \put(20,0){\makebox(30,10){\eqref{sym3-1p}}}
725: \put(50,5){\makebox(30,10){$\scriptstyle s_2$}}
726: \put(50,5){\vector(1,0){30}}
727: \put(80,5){\vector(-1,0){30}}
728: \put(80,0){\makebox(30,10){\eqref{sym3-2p}}}
729: \put(110,5){\makebox(30,10){$\scriptstyle s_0$}}
730: \put(110,5){\vector(1,0){30}}
731: \put(140,5){\vector(-1,0){30}}
732: \put(140,0){\makebox(30,10){\eqref{sym3-3p}}}
733: \put(175,5){\oval(8,8)[r]}
734: \put(175,0){\vector(-1,0){4}}
735: \put(175,9){\vector(-1,0){4}}
736: \put(180,0){\makebox(10,10){$\scriptstyle s_2$}}
737: \end{picture}
738: \end{center}
739: \end{proposition}
740:
741: \bigskip
742:
743: In section \ref{MS3}, we will show these symmetric solutions are
744: monodromy solvable.
745:
746: \subsection{Comparison with classical solutions}
747: In the case of $\alpha_{0}=0\ (\delta=0)$, the sixth Painlev\'e
748: equation \eqref{P6H} admits the Riccati type solution
749: \begin{align}
750: \begin{aligned}
751: y&=t,\\
752: %=\frac 12 + \left(t-\frac 12\right),\\
753: t(t-1)\frac{dz}{dt}&=-t(t-1)z^{2}+\Bigl\{1-2t+
754: \alpha_{3}t+\alpha_{4}(t-1)
755: \Bigr\}z-(\alpha_{1}+\alpha_{2})\alpha_{2}.
756: \end{aligned}
757: \label{Riccati}
758: \end{align}
759: If $\alpha_{3}=\alpha_{4}$, \eqref{Riccati} admits a solution
760: \begin{align*}
761: y&=1/2 + \left(t-1/2\right),\\
762: z&=\frac{d}{dt}
763: \log
764: \left[_{2}F_{1}\left(
765: \alpha_{2}/2,\,
766: (\alpha_{1}+\alpha_{2})/2,\,
767: 1/2,\,
768: 4(t-1/2)^2\right)
769: \right],
770: \end{align*}
771: which is a special case of \eqref{sym2-1}.
772: If $\alpha_{1}=\alpha_{3}=\alpha_{4}$, \eqref{Riccati} admits a solution
773: \begin{align*}
774: y=&-\omega^2+\left(t+\omega^{2}\right),\\
775: z=&\frac{d}{dt}
776: \log
777: \left[
778: _{2}F_{1}\left(\frac{1+\alpha_{1}}2,\,
779: \frac{1+3\alpha_{1}}6,\,
780: \frac 23,\,
781: -\left(\frac{-\omega t-1}{t+\omega}\right)^3
782: \right)
783: \right]\\
784: &\qquad \qquad
785: +\frac{t(1-t)(t+\omega^2)+(t^3+3\omega t^{2}+3(1-\omega)t-2)\alpha_{1}}
786: {2t(t-1)(t^2-t+1)},
787: \end{align*}
788: which is a special case of \eqref{sym3-2}.
789:
790:
791: \section{Transformation of the Linearization}
792:
793: Since the Painlev\'e equation is the isomonodromic deformation condition,
794: to show the monodromy solvability of a solution of Painlev\'e equation,
795: it is sufficient to show the monodromy solvability at a special $t=t_{0}$.
796: In the following, we will set such special $t=t_{0}$ at the fixed points
797: of $\sigma_{1}$ and $\sigma_{2}\circ\sigma_{1}$.
798:
799: \subsection{Monodromy Solvability of the Symmetric Solution of $\sigma_{1}$}
800: \label{MS2}
801: We substitute the symmetric solution \eqref{sym2-1}
802: into \eqref{OG-1} and take the limit $t\to 1/2$, then
803: \begin{align}
804: \frac{d^2 \psi}{d x^2}+
805: \left(
806: \frac{1-\alpha_{3}}{x}+\frac{1-\alpha_{3}}{x-1}+\frac{-\alpha_{0}}{x-1/
807: 2}
808: \right)\frac{d \psi}{d x}+
809: \frac{(\alpha_{1}+\alpha_{2})\alpha_{2}}{x(x-1)}
810: \psi=0.\label{le2}
811: \end{align}
812: This is Heun's equation, whose Riemann scheme is
813: \[
814: P\left\{
815: \begin{matrix}
816: 0 & 1 & 1/2 & \infty \\
817: 0 & 0 & 0 & \alpha_{2} \\
818: \alpha_{3} & \alpha_{3} & \alpha_{0}+1 & \alpha_{1}+\alpha_{2}
819: \end{matrix}; x
820: \right\}.
821: \]
822: Now we set $(2x-1)^2=1-\xi$, then \eqref{le2} becomes
823: \[
824: \frac{d^2 \psi}{d \xi^2}+
825: \left(
826: \frac{1-\alpha_{3}}{\xi}+\frac{1-\alpha_{0}}{2(\xi-1)}
827: \right)\frac{d \psi}{d \xi}+
828: \frac{(\alpha_{1}+\alpha_{2})\alpha_{2}}{4\xi(\xi-1)}
829: \psi=0,
830: \]
831: which is the hypergeometric equation.
832:
833: %Then we have the following theorem.
834: \begin{theorem}
835: The symmetric solution \eqref{sym2-1} is monodromy
836: solvable. Its linearization \eqref{OG-1} at $t=1/2$ is reduced to the
837: hyper\-geometric equation:
838: \[
839: P\left\{
840: \begin{matrix}
841: 0 & 1 & \infty \\
842: 0 & 0 & \frac{\alpha_{2}}{2}\\
843: \alpha_{3}&\frac{\alpha_{0}+1}{2}&\frac{\alpha_{1}+\alpha_{2}}{2}
844: \end{matrix}; \xi
845: \right\}.
846: \]
847: A fundamental solution of \eqref{le2} is given by
848: \[
849: \left({}_{2}F_{1}\left(
850: \frac{\alpha_{2}}2, \frac{\alpha_{1}+\alpha_{2}}2,
851: 1-\alpha_{3}; 4x(1-x)
852: \right),\
853: (4x(1-x))^{\alpha_{3}}{}_{2}F_{1}\left(
854: \frac{\alpha_{2}}2+\alpha_{3},
855: \frac{\alpha_{1}+\alpha_{2}}2+\alpha_{3}, 1+\alpha_{3};4x(1-x)
856: \right)
857: \right).
858: \]
859: \end{theorem}
860:
861: \bigskip
862:
863: In the same way, the linearizations of \eqref{sym2-2}, \eqref{sym2-3}
864: and \eqref{sym2-4} can be reduced to the hypergeometric equation.
865:
866: In section \ref{LM2}, we will calculate the linear monodromy of the
867: symmetric solutions of~$\sigma_{1}$ explicitly.
868:
869: \subsection{Monodromy Solvability of the Symmetric Solution of
870: $\sigma_{2}\circ\sigma_{1}$ }\label{MS3}
871: We substitute the symmetric solution \eqref{sym3-2} into \eqref{OG-1}
872: and take the limit $t\to -\omega^2$, then we have
873: \begin{align}
874: \frac{d^2 \psi}{d x^2}+
875: \left(
876: \frac{1-\alpha_{1}}x+\frac{1-\alpha_{1}}{x-1}+
877: \frac{-\alpha_{0}}{x+\omega^2}
878: \right)\frac{d \psi}{d x}+
879: \left(
880: \frac{\alpha_{2}(\alpha_{1}+\alpha_{2})}{x(x-1)}-
881: \frac{2\omega+1}3\frac{\alpha_{0}\alpha_{2}}{x(x-1)(x+\omega^2)}
882: \right)
883: \psi=0.\label{le3}
884: \end{align}
885: This is Heun's equation, whose Riemann scheme is
886: \[
887: P\left\{
888: \begin{matrix}
889: 0 & 1 & -\omega^2 & \infty \\
890: 0 & 0 & 0 & \alpha_{2} \\
891: \alpha_{1} & \alpha_{1} & 1+\alpha_{0} & \alpha_{1}+\alpha_{2}
892: \end{matrix}; x
893: \right\}
894: =(x+\omega^2)^{-\alpha_{2}}P\left\{
895: \begin{matrix}
896: 0&1&-\omega^2&\infty\\
897: 0&0&\alpha_{2}&0\\
898: \alpha_{1}&\alpha_{1}&1+\alpha_{0}+\alpha_{2}&\alpha_{1}
899: \end{matrix}; x
900: \right\}.
901: \]
902: By the transformation
903: \[
904: \psi=(x+\omega^2)^{-\alpha_{2}}\tilde{\psi},\qquad
905: \xi=\omega\frac{x+\omega}{x+\omega^2},\qquad
906: \eta=\xi^3,
907: \]
908: \eqref{le3} is reduced to the hyper\-geometric equation
909: \begin{align*}
910: \eta(\eta-1)\frac{d^2\tilde{\psi}}{d\eta^2}+
911: \left\{\frac 23-\left(1+\frac{1+\alpha_{0}+2\alpha_{2}}3\right)\eta\right\}
912: \frac{d\tilde{\psi}}{d\eta}-
913: \frac{\alpha_{2}}9\left(
914: 1+\alpha_{0}+\alpha_{2}
915: \right)\tilde{\psi}=0.%\label{le32}
916: \end{align*}
917: Then we have the following.
918:
919: \begin{theorem}
920: The symmetric solution \eqref{sym3-2} on $\sigma_{2}\circ\sigma_{1}$ is
921: monodromy solvable. Its linearization
922: \eqref{OG-1} at $t=-\omega^2$ is reduced to the hypergeometric equation
923: \[
924: P\left\{
925: \begin{matrix}
926: 0&1&\infty\\
927: 0&0&\alpha_{2}/3\\
928: 1/3&\alpha_{1}&(1+\alpha_{0}+\alpha_{2})/3
929: \end{matrix};\eta
930: \right\}.
931: \]
932: A fundamental solution is
933: \[
934: \left(
935: {}_{2}F_{1}\left(\frac{\alpha_{2}}{3},\frac{1}{3}(\alpha_{0}+\alpha_{2}+
936: 1), \frac{2}{3};\eta
937: \right),\ \\
938: \eta^{1/3}{}_{2}F_{1}\left(
939: \frac{\alpha_{2}+1}{3},\frac 13(\alpha_{0}+\alpha_{2}+2), \frac 43 ; \eta
940: \right)
941: \right).
942: \]
943: \end{theorem}
944:
945: \bigskip
946:
947: In the same way, \eqref{sym3-1} and \eqref{sym3-3} can be reduced to the
948: hypergeometric equation. %Therefore, we can calculate their monodromy.
949:
950: In section \ref{LM3}, we will calculate the linear monodromy of the
951: symmetric solution of $\sigma_{2}\circ\sigma_{1}$ explicitly.
952:
953: \section{Linear Monodromy}
954: %In the following,
955: %we take paths $\gamma_{j}$ around singular points $a_{j}$ as a standard
956: %way in the figure.
957: %\begin{center}
958: %\includegraphics{path.eps}
959: %\end{center}
960: %We represent a choice of paths corresponding to the monodromy matrices
961: %$M_{j}$ around the singular points $a_{j}$ as
962: %\[
963: % \left[\begin{matrix}
964: %a_{1} & a_{2} & \cdots & a_{k}\\
965: %M_{1} & M_{2} & \cdots & M_{k}
966: %\end{matrix}
967: %\right],
968: %\]
969: %where the fundamental solution $\Psi=(\psi_{1},\psi_{2})$ is
970: %analytically continued along
971: %$\gamma_{j}$ as
972: %\[
973: % \Psi(\gamma_{j}(x))=\Psi(x)M_{j}.
974: %\]
975: %The monodromy matrices satisfy the following relation
976: %\[
977: % M_{k}\cdots M_{2}M_{1}=1.
978: %\]
979:
980:
981: \subsection{Linear Monodromy of the Symmetric Solution of $\sigma_{1}$}
982: \label{LM2}
983:
984: By $(2x-1)^2=1-\xi$,
985: $\xi$-surface $\mathbb{P}_{1}\setminus\{0,1,\infty\}$ is double covered
986: by the $x$-surface
987: $\mathbb{P}_{1}\setminus\{0,1/2,1,\infty\}$.
988: Let us make a cut from $\xi=1$ to $\infty$ ($x=1/2$ to $\infty$), and
989: suppose that $x=1$ is on the second $\xi$-plane, and the base point and
990: $x=0$ are on the first $\xi$-plane.
991:
992: Let us denote the standard paths around
993: $x=0,1/2,1,\infty$ by
994: $\gamma_{0},\gamma_{1/2},\gamma_{1},\gamma_{\infty}$,
995: and denote the standard paths around $\xi=0,1,\infty$ by
996: $L_{0},L_{1},L_{\infty}$.
997:
998: \begin{center}
999: \includegraphics[width=12cm]{s001.eps}\\
1000: \includegraphics[width=12cm]{s002.eps}
1001: \end{center}
1002:
1003: Then we have
1004: $$
1005: \gamma_{0}=L_{0}, \qquad
1006: \gamma_{1/2}=L_{1}^2, \qquad
1007: \gamma_{1}=L_{1}^{-1}L_{0}L_{1}, \qquad
1008: \gamma_{\infty}=L_{\infty}^2,
1009: $$
1010: which satisfy
1011: \[
1012: \gamma_{0}\gamma_{1/2}\gamma_{1}\gamma_{\infty}=
1013: L_{0}(L_{1}^2)(L_{1}^{-1}L_{0}L_{1})L_{\infty}^2=1.
1014: \]
1015: %Therefore,
1016: %for a fundamental solution
1017: %$\Psi=(\psi_{1},\psi_{2})$ of \eqref{OG-1},
1018: %their monodromy is represented by
1019: %$$
1020: %M_{0}=G_{0},\qquad
1021: %M_{t}=G_{1}^2, \qquad
1022: %M_{1}=G_{1}G_{0}G_{1}^{-1}, \qquad
1023: %M_{\infty}=G_{\infty}^2,
1024: %$$
1025: %which satisfies
1026: %\[
1027: % M_{\infty}M_{1}M_{t}M_{0}=
1028: %G_{\infty}^{2}(G_{1}G_{0}G_{1}^{-1})G_{1}^2G_{0}=1.
1029: %\]
1030:
1031: Then we obtain monodromy matrix $M_{j}$ along path $\gamma_{j}$
1032: ($j=0, 1/2, 1, \infty$).
1033: \begin{theorem}\label{mon2-thm}
1034: The linear monodromy of the symmetric solution \eqref{sym2} of $\sigma_{1}$
1035: is represented as
1036: \begin{gather*}
1037: M_{0}=\Gamma_{{0\infty}}\Lambda_{0}\Gamma_{0\infty}^{-1},\qquad
1038: M_{1/2}=\Gamma_{{\frac 12\infty}}
1039: \Lambda_{\frac 12}^2\Gamma_{\frac 12\infty}^{-1},\\
1040: M_{1}=(\Gamma_{{\frac 12\infty}}\Lambda_{\frac 12}\Gamma_{\frac 12\infty}^{-1})
1041: (\Gamma_{{0\infty}}\Lambda_{1}\Gamma_{0\infty}^{-1})
1042: (\Gamma_{{\frac 12\infty}}
1043: \Lambda_{\frac 12}^{-1}\Gamma_{\frac 12\infty}^{-1}),\qquad
1044: M_{\infty}=(e^{2\pi i T_{\infty}})^2,
1045: \end{gather*}
1046: where
1047: \begin{gather*}
1048: \Lambda_{0}=\Lambda_{1}=\begin{pmatrix}
1049: 1&0\\0&e^{2\pi i \alpha_{3}}
1050: \end{pmatrix},\qquad
1051: \Lambda_{\frac 12}=\begin{pmatrix}
1052: 1&0\\
1053: 0&-e^{\pi i \alpha_{0}}
1054: \end{pmatrix},\qquad
1055: e^{2\pi i T_{\infty}}=\begin{pmatrix}
1056: e^{\pi i\alpha_{2}}&0\\0&e^{\pi i (\alpha_{1}+\alpha_{2})}
1057: \end{pmatrix},\\
1058: \Gamma_{0\infty}=\Gamma_{1\infty}=\begin{pmatrix}
1059: e^{-\alpha_{2}\pi i/2}
1060: \frac{\Gamma(1-\alpha_{3})\Gamma(\alpha_{1}/2)}
1061: {\Gamma((\alpha_{1}+\alpha_{2})/2)\Gamma(1-\alpha_3-\alpha_{2}/2)}
1062: & %%%%%%%%%%%%%%%%%%%%%%%%%
1063: e^{-(\alpha_{3}+\alpha_{2}/2)\pi i}\frac
1064: {\Gamma(1+\alpha_{3})\Gamma(\alpha_{1}/2)}
1065: {\Gamma(1-\alpha_{2}/2)\Gamma((1-\alpha_{0}-\alpha_{2})/2)}
1066: \\[+5pt] %%%%%%%%%%%%%%%%%%%%%%%%
1067: e^{-(\alpha_{1}+\alpha_{2})\pi i/2}\frac
1068: {\Gamma(1-\alpha_{3})\Gamma(-\alpha_{1}/2)}
1069: {\Gamma(\alpha_{2}/2)\Gamma((1+\alpha_{0}+\alpha_{2})/2)}
1070: & %%%%%%%%%%%%%%%%%%%%%%%%%%%
1071: e^{(\alpha_{0}+\alpha_{2}-1)\pi i/2}\frac
1072: {\Gamma(1+\alpha_{3})\Gamma(-\alpha_{1}/2)}
1073: {\Gamma((1-\alpha_{1}-\alpha_{2})/2)\Gamma(\alpha_{3}+\alpha_{2}/2)}
1074: \end{pmatrix},
1075: \\ %%%%%%%%%%%%%%%%%%%%%%%%%%%%%
1076: \Gamma_{\frac 12 \infty}=\begin{pmatrix}
1077: \frac{\Gamma((3+\alpha_{0})/2)\Gamma(-\alpha_{1}/2)}
1078: {\Gamma((1+\alpha_{0}+\alpha_{2})/2)\Gamma((2-\alpha_{1}-\alpha_{2})/2)}
1079: & %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
1080: -\frac{\Gamma((3+\alpha_{0})/2)\Gamma(\alpha_{1}/2)}
1081: {\Gamma(1-\alpha_{3}-\alpha_{2}/2)\Gamma(1-\alpha_{2}/2)}
1082: \\[+5pt] %%%%%%%%%%%%%%%%%%%%%%%%%
1083: -e^{\pi i ((1+\alpha_{0})/2)}\frac
1084: {\Gamma((1-\alpha_{0})/2)\Gamma(\alpha_{1}/2)}
1085: {\Gamma(\alpha_{2}/2)\Gamma(\alpha_{3}+\alpha_{2}/2)}
1086: & %%%%%%%%%%%%%
1087: e^{\pi i ((1+\alpha_{0})/2)}\frac
1088: {\Gamma((1-\alpha_{0})/2)\Gamma(\alpha_{1}/2)}
1089: {\Gamma((\alpha_{1}+\alpha_{2})/2)\Gamma((1-\alpha_{0}-\alpha_{2})/2)}
1090: \end{pmatrix}.
1091: \end{gather*}
1092: \end{theorem}
1093:
1094: \subsection{Linear Monodromy of the Symmetric Solution of
1095: $\sigma_{2}\circ\sigma_{1}$} \label{LM3}
1096:
1097: Let us denote the standard paths around $\xi=\omega, \omega^2, 1,
1098: \infty$ by $\gamma_{\omega}$, $\gamma_{\omega^2}$, $\gamma_{1}$,
1099: $\gamma_{\infty}$, and the standard paths around $\eta=0, 1, \infty$ by
1100: $L_{0}$, $L_{1}$, $L_{\infty}$.% We set on $x$-space
1101: %\[
1102: % \begin{bmatrix}
1103: %0 & t & 1 & \infty\\
1104: %M_{0}&M_{t}&M_{1}&M_{\infty}
1105: %\end{bmatrix},
1106: %\]
1107: %on $\xi$-space
1108: %\[
1109: % \begin{bmatrix}
1110: % \omega & \omega^2 & 1 & \infty\\
1111: % \tilde{M}_{\omega} & \tilde{M}_{\omega^2} & \tilde{M}_{1} & \tilde{M}_{\infty}
1112: %\end{bmatrix},
1113: %\]
1114: %and on $\eta$-space
1115: %\[
1116: % \begin{bmatrix}
1117: %0&1&\infty\\
1118: %G_{0}&G_{1}&G_{\infty}
1119: %\end{bmatrix}.
1120: %\]
1121: %Then
1122: %\[
1123: % M_{0}=\tilde{M}_{1}
1124: %\]
1125: %%\begin{center}
1126: %%\includegraphics{xi30.eps} \hspace{10mm}
1127: %%\includegraphics{eta30.eps}
1128: %%\end{center}
1129:
1130: The $\xi$-space and the $\eta$-space is connected by $\xi^3=\eta$.
1131:
1132: \begin{center}
1133: \includegraphics{xi31.eps}\hspace{10mm}
1134: \includegraphics{eta31.eps}
1135: \end{center}
1136:
1137: \begin{center}
1138: \includegraphics{xi32.eps}\hspace{10mm}
1139: \includegraphics{eta32.eps}
1140: \end{center}
1141:
1142: Then we have
1143: \[
1144: \gamma_{\omega}=L_{0}L_{1}L_{0}^{-1},\qquad
1145: \gamma_{\omega^{2}}=L_{0}^{-1}L_{1}L_{0},\qquad
1146: \gamma_{1}=L_{1},\qquad
1147: \gamma_{\infty}=L_{\infty}^3,
1148: \]
1149: where
1150: \[
1151: \gamma_{\omega}\gamma_{\omega^2}\gamma_{1}\gamma_{\infty}=
1152: (L_{0}L_{1}L_{0}^{-1})(L_{0}^{-1}L_{1}L_{0})L_{1}(L_{\infty}^3)
1153: =L_{0}L_{1}L_{0}^{-3}L_{\infty}.
1154: \]
1155: Since the monodromy along the path $L_{0}^{3}$ is $1$, the monodromy
1156: along the path
1157: $\gamma_{\omega}\gamma_{\omega^2}\gamma_{1}\gamma_{\infty}$ is $1$.
1158:
1159:
1160: Then we obtain monodromy matrix $M_{j}$ along path $\gamma_{j}$
1161: ($j=\omega, \omega^2, 1, \infty$).
1162:
1163: \begin{theorem} \label{mon3-thm}
1164: The linear monodromy of the symmetric solution \eqref{sym3-2} of $\sigma_{2}\circ\sigma_{1}$
1165: is represented as follows:
1166: \begin{gather*}
1167: M_{\omega^2}=e^{-2\pi i \alpha_{2}}\left(
1168: e^{2\pi i T_{\infty}}
1169: \right)^{3},\qquad
1170: M_{0}=(\Gamma_{1\infty}\Lambda_{1}\Gamma_{1\infty}^{-1}),
1171: \\
1172: M_{1}=(\Gamma_{0\infty}\Lambda_{0}\Gamma_{0\infty}^{-1})
1173: (\Gamma_{1\infty}\Lambda_{1}\Gamma_{1\infty}^{-1})
1174: (\Gamma_{0\infty}\Lambda_{0}^{-1}\Gamma_{0\infty}^{-1}),\\
1175: M_{\infty}=e^{2\pi i\alpha_{2}}
1176: (\Gamma_{0\infty}\Lambda_{0}^{-1}\Gamma_{0\infty}^{-1})
1177: (\Gamma_{1\infty}\Lambda_{1}\Gamma_{1\infty}^{-1})
1178: (\Gamma_{0\infty}\Lambda_{0}\Gamma_{0\infty}^{-1}),
1179: \end{gather*}
1180: where
1181: \begin{gather*}
1182: \Lambda_{0}=\begin{pmatrix}1&0\\0&e^{2\pi i/3}\end{pmatrix},\qquad
1183: \Lambda_{1}=\begin{pmatrix}1&0\\0&e^{2\pi i
1184: (1-\alpha_{0}-2\alpha_{2})/3}\end{pmatrix},\qquad
1185: e^{2\pi i T_{\infty}}=\begin{pmatrix}
1186: e^{2\pi i\alpha_{2}/3}&0\\0&e^{2\pi i(1+\alpha_{0}+\alpha_{2})/3}
1187: \end{pmatrix},\\ %%%%%%%%%%%%%%%%%%%%%%
1188: \Gamma_{0\infty}=\begin{pmatrix}
1189: e^{\alpha_{2}\pi i/3}\frac{\Gamma(2/3)\Gamma((1+\alpha_{0})/3)}
1190: {\Gamma((1+\alpha_{0}+\alpha_{2})/3)\Gamma((2-\alpha_{2})/3)}
1191: & %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
1192: e^{(1+\alpha_{2})\pi i/3}\frac
1193: {\Gamma(4/3)\Gamma((1+\alpha_{0})/3)}
1194: {\Gamma(1-\alpha_{2}/3)\Gamma((2+\alpha_{0}+\alpha_{2})/3)}
1195: \\[+5pt] %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
1196: e^{(1+\alpha_{0}+\alpha_{2})\pi i/3}\frac
1197: {\Gamma(2/3)\Gamma((1+\alpha_{0})/(-3))}
1198: {\Gamma(\alpha_{2}/3)\Gamma((1-\alpha_{0}-\alpha_{2})/3)}
1199: & %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
1200: e^{(2+\alpha_{0}+\alpha_{2})\pi i/3}\frac
1201: {\Gamma(4/3)\Gamma((1+\alpha_{0})/(-3))}
1202: {\Gamma((2-\alpha_{0}-\alpha_{2})/3)\Gamma((1+\alpha_{2})/3)}
1203: \end{pmatrix},\\ %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
1204: \Gamma_{1\infty}=\begin{pmatrix}
1205: \frac{\Gamma((2+\alpha_{0}+2\alpha_{2})/3)\Gamma((1+\alpha_{0})/3)}
1206: {\Gamma((1+\alpha_{0}+\alpha_{2})/3)\Gamma((2+\alpha_{0}+\alpha_{2})/3)}
1207: & %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
1208: e^{(-1+\alpha_{0}+2\alpha_{2})\pi i/3}
1209: \frac{\Gamma((4-\alpha_{0}-2\alpha_{2})/3)\Gamma((1+\alpha_{0})/3)}
1210: {\Gamma((2-\alpha_{2})/3)\Gamma(1-\alpha_{2}/3)}
1211: \\[+5pt] %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
1212: \frac{\Gamma((2+\alpha_{0}+2\alpha_{2})/3)\Gamma((1+\alpha_{0})/(-3))}
1213: {\Gamma(\alpha_{2}/3)\Gamma((1+\alpha_{2})/3)}
1214: & %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
1215: e^{(-1+\alpha_{0}+2\alpha_{2})\pi i/3}\frac
1216: {\Gamma((4-\alpha_{0}-2\alpha_{2})/3)\Gamma((1+\alpha_{0})/(-3))}
1217: {\Gamma((1-\alpha_{0}-\alpha_{2})/3)\Gamma((2-\alpha_{0}-\alpha_{2})/3)}
1218: \end{pmatrix}.
1219: \end{gather*}
1220: \end{theorem}
1221:
1222:
1223:
1224:
1225:
1226:
1227: \section{Characterizations of the monodromy}\label{char}
1228:
1229: In this section, we will characterize the linear monodromy of our
1230: symmetric solutions on Fricke's cubic surface of monodromy.
1231:
1232: \bigskip
1233:
1234: To normalize the monodromy matrices in $SL(2,\mathbb{C})$,
1235: we replace $\psi$ of the
1236: linearization \eqref{OG-1} as
1237: \[
1238: \psi \to x^{\alpha_{4}/2}(x-1)^{\alpha_{3}/2}(x-t)^{(\alpha_{0}-1)/2}\psi,
1239: \]
1240: then the Riemann scheme of \eqref{OG-1} becomes
1241: \[
1242: P\left\{x;
1243: \begin{matrix}
1244: 0 &1 &t &y &\infty\\
1245: -\frac{\alpha_{4}}2&-\frac{\alpha_{3}}2&-\frac{\alpha_{0}-1}2&0&
1246: -\frac{\alpha_{1}}2\\
1247: \frac{\alpha_{4}}2&\frac{\alpha_{3}}2&\frac{\alpha_{0}-1}2+1&2&
1248: \frac{\alpha_{1}}2
1249: \end{matrix}\right\},
1250: \]
1251: and then the monodromy matrices
1252: $\left\{M_{0},M_{t},M_{1},M_{\infty}\right\}$
1253: become elements of
1254: $SL(2,\mathbb{C})$. Hereafter we use this linearization.
1255:
1256:
1257:
1258: \subsection{Linear monodromy of the Symmetric Solutions of $\sigma_{1}$}
1259: \label{char2}
1260:
1261: From the theorem \ref{mon2-thm}, we can calculate
1262: $\{p_{0}, p_{1}, p_{t}, p_{\infty}; p_{01}, p_{1t}, p_{t0}\}$ for the
1263: symmetric solution \eqref{sym2-1}. We have
1264: \[
1265: p_{0}=p_{1}=2\cos\pi\alpha_{3}, \qquad
1266: p_{t}=2\cos\pi(\alpha_{0}+1), \qquad
1267: p_{\infty}=2\cos\pi\alpha_{1},
1268: \]
1269: and
1270: \begin{align}
1271: &p_{1t}=p_{t0}=2\left(
1272: \cos\pi(\alpha_{3}+1)+\cos\frac{\pi}2(\alpha_{0}-\alpha_{1}-1)
1273: +\cos\frac{\pi}2(\alpha_{0}+\alpha_{1}-1)
1274: \right),\label{X21}\\
1275: &p_{01}=-2\left(1+\cos\pi(\alpha_{0}+1)+\cos\pi\alpha_{1}+
1276: 4\cos\frac\pi 2\alpha_{1}\cos\pi\alpha_{3}\cos\frac\pi 2(\alpha_{0}+1)\right).
1277: \end{align}
1278: In the same way, for the solution \eqref{sym2-2}, we have
1279: \begin{align}
1280: & p_{0}=p_{1}=2\cos\pi\alpha_{3}, \qquad
1281: p_{t}=2\cos\pi(\alpha_{0}+1), \qquad
1282: p_{\infty}=2\cos\pi\alpha_{1},\notag \\
1283: &p_{1t}=p_{t0}=2\left(
1284: \cos\pi(\alpha_{3}+1)+\cos\frac{\pi}2(\alpha_{0}-\alpha_{1}+1)
1285: +\cos\frac{\pi}2(\alpha_{0}+\alpha_{1}+1)
1286: \right),\label{X22}\\
1287: &p_{01}=-2\left(1+\cos\pi(\alpha_{0}-1)+\cos\pi\alpha_{1}+
1288: 4\cos\frac\pi 2\alpha_{1}\cos\pi\alpha_{3}\cos\frac\pi 2(\alpha_{0}-1)\right),
1289: \end{align}
1290: for \eqref{sym2-3}, we have
1291: \begin{align}
1292: & p_{0}=p_{1}=2\cos\pi\alpha_{3}, \qquad
1293: p_{t}=2\cos\pi(\alpha_{0}+1), \qquad
1294: p_{\infty}=2\cos\pi\alpha_{1},\notag \\
1295: &p_{1t}=p_{t0}=2\left(
1296: \cos\pi\alpha_{3}+\cos\frac{\pi}2(\alpha_{0}-\alpha_{1}-1)
1297: +\cos\frac{\pi}2(\alpha_{0}+\alpha_{1}+1)
1298: \right),\label{X23}\\
1299: &p_{01}=-2\left(1+\cos\pi\alpha_{0}+\cos\pi(\alpha_{1}+1)+
1300: 4\cos\frac\pi 2(\alpha_{1}+1)\cos\pi\alpha_{3}\cos\frac\pi 2\alpha_{0}\right),
1301: \end{align}
1302: and, for \eqref{sym2-4}, we have
1303: \begin{align}
1304: & p_{0}=p_{1}=2\cos\pi\alpha_{3}, \qquad
1305: p_{t}=2\cos\pi(\alpha_{0}+1), \qquad
1306: p_{\infty}=2\cos\pi\alpha_{1},\notag \\
1307: &p_{1t}=p_{t0}=2\left(
1308: \cos\pi\alpha_{3}+\cos\frac{\pi}2(\alpha_{0}-\alpha_{1}+1)
1309: +\cos\frac{\pi}2(\alpha_{0}+\alpha_{1}-1)
1310: \right),\label{X24}\\
1311: &p_{01}=-2\left(1+\cos\pi\alpha_{0}+\cos\pi(\alpha_{1}-1)+
1312: 4\cos\frac\pi 2(\alpha_{1}-1)\cos\pi\alpha_{3}\cos\frac\pi 2\alpha_{0}\right).
1313: \end{align}
1314: Certainly, they satisfies \eqref{eq-mon}.
1315:
1316: \bigskip
1317:
1318: Conversely, if we set
1319: \begin{gather*}
1320: X= p_{1t}=p_{t0}, \qquad \qquad Y= p_{01},\\
1321: A=p_{0}=p_{1}=2\cos\pi\alpha_{3}, \qquad
1322: B=p_{t}=2\cos\pi(\alpha_{0}+1), \qquad
1323: C=p_{\infty}=2\cos\pi\alpha_{1},
1324: \end{gather*}
1325: then \eqref{eq-mon} becomes
1326: \[
1327: X^2Y+Y^2+2X^2-(A^2+BC)Y-2(AB+AC)X+2A^2+B^2+C^2+A^2BC-4=0.
1328: \]
1329: As an equation of $Y$, the discriminant of this equation can be
1330: factorized as follows
1331: \[
1332: (X-X_{1})(X-X_{2})(X-X_{3})(X-X_{4}),
1333: \]
1334: where $X_{1}$, $X_{2}$, $X_{3}$ and $X_{4}$ are the right hand sides of
1335: \eqref{X21}, \eqref{X22}, \eqref{X23} and \eqref{X24}.
1336:
1337: Now, we have the following.
1338: \begin{theorem}
1339: Under the constraints $p_{0}=p_{1}$ and $p_{0t}=p_{t1}$, the relation of
1340: monodromy \eqref{eq-mon} admits a double root as an equation of $p_{01}$,
1341: if and only if
1342: $\left\{M_{0}, M_{t}, M_{1}, M_{\infty}\right\}$ is the linear monodromy
1343: of a symmetric solution of $\sigma_{1}$.
1344: \end{theorem}
1345:
1346: \subsection{Linear monodromy of the Symmetric Solutions of $\sigma_{2}\circ\sigma_{1}$}
1347: \label{char3}
1348:
1349: From the theorem \ref{mon3-thm}, we can calculate
1350: $\{p_{0}, p_{1}, p_{t}, p_{\infty}; p_{01}, p_{1t}, p_{t0}\}$ for the
1351: symmetric solution \eqref{sym3-2}. We have
1352: \begin{align}
1353: & p_{0}=p_{1}=p_{\infty}=2\cos(\pi\alpha_{1}),\qquad
1354: p_{t}=2\cos\pi(\alpha_{0}-1),\\
1355: &p_{0t}=p_{1t}=p_{t\infty}=-1-2\cos\left(\tfrac{2\pi}3\alpha_{0}\right)
1356: +4\cos\left(\tfrac\pi 3 \alpha_{0}\right)\cos\pi\alpha_{1},
1357: \label{X31}
1358: \end{align}
1359: In the same way, for the solution \eqref{sym3-1}, we have
1360: \begin{align}
1361: & p_{0}=p_{1}=p_{\infty}=2\cos(\pi\alpha_{1}),\qquad
1362: p_{t}=2\cos\pi(\alpha_{0}-1),\\
1363: &p_{0t}=p_{1t}=p_{t\infty}=-1-2\cos\left(\tfrac{2\pi}3(\alpha_{0}-2)\right)
1364: +4\cos\left(\tfrac\pi 3 (\alpha_{0}-2)\right)\cos\pi\alpha_{1},
1365: \label{X32}
1366: \end{align}
1367: and, for \eqref{sym3-3}, we have
1368: \begin{align}
1369: & p_{0}=p_{1}=p_{\infty}=2\cos(\pi\alpha_{1}),\qquad
1370: p_{t}=2\cos\pi(\alpha_{0}-1),\\
1371: &p_{0t}=p_{1t}=p_{t\infty}=-1-2\cos\left(\tfrac{2\pi}3(\alpha_{0}+2)\right)
1372: +4\cos\left(\tfrac\pi 3 (\alpha_{0}+2)\right)\cos\pi\alpha_{1}.
1373: \label{X33}
1374: \end{align}
1375:
1376: Now,
1377: $\{p_{0}, p_{1}, p_{t}, p_{\infty}; p_{\infty 1}, p_{t1}, p_{t \infty}\}$
1378: also parameterize the monodromy, and they satisfy the following relation
1379: \begin{multline}
1380: p_{\infty 1}p_{1t}p_{t\infty}+p_{\infty 1}^2+p_{1 t}^2+p_{t\infty}^2
1381: -(p_{\infty}p_{1}+p_{t}p_{0})
1382: p_{\infty
1383: 1}-(p_{1}p_{t}+p_{\infty}p_{0})p_{1t}-(p_{t}p_{\infty}+p_{1}p_{0})p_
1384: {t\infty}
1385: \\
1386: +p_{\infty}^2+p_{1}^2+p_{t}^2+p_{0}^{2}
1387: +p_{0}p_{t}p_{1}p_{t}p_{\infty}-4=0.
1388: \label{eq-mon2}
1389: \end{multline}
1390: Since $M_{0}$, $M_{t}$, $M_{1}$, $M_{\infty}$ are elements of
1391: $SL(2,\mathbb{C})$, we have
1392: \[
1393: p_{0t}=\mathrm{tr}M_{0}M_{t}=\mathrm{tr}\left(M_{1}M_{\infty}\right)^{-1}=
1394: \mathrm{tr}M_{\infty}M_{1}=p_{\infty 1}.
1395: \]
1396: And thus,
1397: $\{p_{0}, p_{1}, p_{t}, p_{\infty}; p_{t 0}, p_{t 1}, p_{t \infty}\}$
1398: also parameterize the monodromy.
1399:
1400: \bigskip
1401:
1402: Conversely, if we set
1403: \[
1404: X=p_{t0}=p_{t1}=p_{t\infty}, \qquad
1405: A=p_{0}=p_{1}=p_{\infty}=2\cos(\pi\alpha_{1}), \qquad
1406: B= p_{t}=2\cos\pi(\alpha_{0}-1),
1407: \]
1408: then \eqref{eq-mon2} becomes
1409: \[
1410: X^3+3X^2-3(A^2+AB)X+3A^2+B^2+A^3B-4=0.
1411: \]
1412: We can factorize the left hand side of this equation as follows
1413: \[
1414: (X-X_{1})(X-X_{2})(X-X_{3}),
1415: \]
1416: where $X_{1}$, $X_{2}$, $X_{3}$ are the right hand sides of
1417: \eqref{X31}, \eqref{X32}, \eqref{X33}.
1418:
1419: Now we have the following theorem.
1420: \begin{theorem}
1421: Under the constraints $p_{0}=p_{1}=p_{\infty}$ and
1422: $p_{t0}=p_{t1}=p_{t\infty}$, the relation \eqref{eq-mon2} becomes a
1423: third order equation of $p_{t0}$, whose solutions correspond to
1424: the linear monodromy of the symmetric solutions of
1425: $\sigma_{2}\circ\sigma_{1}$.
1426: \end{theorem}
1427:
1428:
1429:
1430: \begin{thebibliography}{3}
1431:
1432: \bibitem{AK} F.~V.~Andreev, and A.~V.~Kitaev,
1433: Transformations ${RS}_4^2(3)$ of the Ranks $\leq4$ and
1434: Algebraic Solutions of the Sixth Painlev\'e Equation,
1435: { Comm. Math. Phys.} {\bf 228} (2002), 151--176.
1436:
1437: \bibitem{Boa} P.~Boalch,
1438: From Klein to Painleve via Fourier, Laplace and Jimbo,
1439: Proc. Lond. Math. Soc., {\bf 90} (2005), 167-208
1440:
1441: \bibitem {RF1} R.~Fuchs, \"Uber lineare homogene
1442: Differentialgleichungen zweiter Ordnung
1443: mit drei im endlichen gelegenen wesentlich singul\"aren Stellen,
1444: Math. Ann., $\boldsymbol{63}$
1445: (1906), 301--321.
1446:
1447: \bibitem {RF2} R.~Fuchs, \"Uber lineare homogene Differentialgleichungen zweiter Ordnung
1448: mit drei im endlichen gelegenen wesentlich singul\"aren Stellen, Math. Ann., $\boldsymbol{70}$
1449: (1911), 525--549.
1450:
1451: \bibitem{Gar} R.~Garnier,
1452: Sur des \'equations diff\'erentielles du troisi\`eme ordre dont
1453: l'int\'egrale g\'en\'erale est uniforme et sur une classe d'\'equations
1454: nouvelles d'ordre sup\'erieur dont l'int\'egrale g\'en\'erale a ses points
1455: critiques fixes,
1456: { Ann. Sci. Ecole Norm. Sup.} Ser. 3, {\bf 29} (1912), 1--126
1457:
1458: \bibitem {Iwa} K.~Iwasaki, An area-preserving action of the modular
1459: group on cubic surfaces and the Painlev\'e VI equation,
1460: Comm. Math. Phys., \textbf{242}, (2003), 185--219
1461:
1462: \bibitem {Jim} M.~Jimbo, Monodromy problem and the boundary condition
1463: for some Painlev\'e equations,
1464: Publ. RIMS, $\boldsymbol{18}$, (1982), 1137--1161
1465:
1466: \bibitem {KK1} K.~Kaneko, A new solution of the fourth Painlev\'e
1467: equation with a solvable monodromy,
1468: Proc. Japan Acad., $\boldsymbol{81}$, Ser. A (2005), 75--79.
1469:
1470: %\bibitem{KK2} K.~Kaneko and Y.~Ohyama, Fifth Painlev\'e transcendents
1471: % which are analytic at the origin, to appear.
1472:
1473: \bibitem {AVK} A.~V.~Kitaev, Symmetric solutions for the first and second Painlev\'e equations,
1474: Zap. Nauchn. Sem. LOMI Steklov. $\boldsymbol{187}$ (1991), 129--138;
1475: \textit{translation in}
1476: J.~Math.~Sci., $\boldsymbol{73}$ (1995), 494--499.
1477:
1478: \bibitem {Fri} F.~Klein and R.~Fricke,
1479: Vorlesungen \"uber die Theorie der automorphen Functionen,
1480: Erster Band, (1897) Teubner.
1481:
1482: \bibitem {MMA} M.~Mazzocco, Picard and Chazy solutions to the Painlev\'e VI equation, Math.
1483: Ann., $\boldsymbol{321}$ (2001), 157--195.
1484:
1485: \bibitem{NY} M.~Noumi and Y.~Yamada, A new Lax pair for the sixth Painlev\'e equation associated with $\hat{\mathfrak{so}}(8)$, in "Microlocal Analysis and Complex Fourier Analysis" (Eds. T.Kawai and K.Fujita), 238-252, World Scientific, 2002
1486:
1487: \bibitem{Oh} Y.~Ohyama and S.~Okumura, R.~Fuchs' problem of the
1488: Painlev\'e equations from the first to the fifth, preprint
1489:
1490: \bibitem {KO} K.~Okamoto, Isomonodromic deformation and Painlev\'e
1491: equations, and the Garnier system. J.~Fac.~Sci.~Univ.~Tokyo
1492: Sect.~1A, Math., $\boldsymbol{33}$ (1986), 575--618.
1493:
1494: \bibitem {EP} E.~Picard, M\'emoire sur la th\'eorie des functions alg\'ebriques de deux variables,
1495: Journal de Liouville, $\boldsymbol{5}$ (1889), 135--319.
1496:
1497: \bibitem {HU1} H.~Umemura, Birational autmorphism groups and differential equations.
1498: Equations differentielles dans le champ complex, Vol.~$\boldsymbol{II}$ (Strasbourg,1985), 119--227,
1499: Publ. Inst. Rech. Math. Av., Univ. Louis Pasteur, Strasbourg.
1500: \end{thebibliography}
1501:
1502: \end{document}@
1503: