1: \documentstyle[cizgi,twoside,paper,12pt]{article}
2: \renewcommand{\thefootnote}{\fnsymbol{footnote}} %* icin \footnote[1]
3: \renewcommand{\theequation}{\arabic{equation}}
4: \renewcommand{\theequation}{\arabic{equation}}
5: \newcommand{\ben}{\begin{enumerate}}
6: \newcommand{\een}{\end{enumerate}}
7: \newcommand{\be}{\begin{equation}}
8: \newcommand{\ee}{\end{equation}}
9: \newcommand{\bea}{\begin{eqnarray}}
10: \newcommand{\eea}{\end{eqnarray}}
11: \newcommand{\bc}{\begin{center}}
12: \newcommand{\ec}{\end{center}}
13: \newcommand{\vs}{\vspace}
14: \newcommand{\mb}{\mbox{\ }}
15: \newcommand{\ra}{\rightarrow}
16: \newcommand{\la}{\leftarrow}
17: \newcommand{\IR}{\mbox{I \hspace{-0.2cm}R}}
18: \newcommand{\IN}{\mbox{I \hspace{-0.2cm}N}}
19: \newcommand{\ol}{\overline}
20: \newcommand{\ul}{\underline}
21: \renewcommand{\baselinestretch}{1.15}
22:
23: \begin{document}
24:
25:
26: \newpage
27: \thispagestyle{empty}
28: \bc
29: {\bf Integrability of
30: Kersten-Krasil'shchik coupled KdV-mKdV equations:
31: singularity analysis and Lax pair}
32: \ec
33: \mb \vs{1cm}
34: \bc
35: Ay\c{s}e Karasu(Kalkanl{\i})$^{*}$, Sergei Yu. Sakovich$^{**}$,\\
36: \'{I}smet Yurdu\c{s}en$^{*}$
37: \ec
38: \mb \vs{1cm}
39: \bc
40: {\em
41: $^{*}$Department of Physics,
42: Middle East Technical University,\\
43: 06531 Ankara, Turkey \\
44: $^{**}$Institute of Physics,
45: National Academy of Sciences,\\
46: 220072 Minsk, Belarus}
47: \ec
48: \vs{1cm}
49: \bc
50: {\bf Abstract}
51: \ec
52: The integrability of a coupled KdV-mKdV system is tested by
53: means of singularity analysis. The true Lax pair associated with
54: this system is obtained by the use of prolongation technique.\\ \\
55: {\bf Keywords}: singularity analysis, prolongation structure, Lax pair\\
56: PACS Codes: 02.90.+p, 02.20.-a, 02.30.Jr\\ \\
57: {\em
58: E-mail: akarasu@metu.edu.tr, saks@pisem.net,
59: ismety@newton.physics.metu.edu.tr}
60:
61: \newpage
62:
63:
64: \setcounter{page}{1}
65: %\baslik{1. Introduction}
66: Very recently, Kersten and Krasil'shchik [1] constructed the
67: recursion operator for symmetries of a coupled KdV-mKdV system
68: \bea
69: u_t&=&-u_{xxx}+6uu_x-3ww_{xxx}-3w_x w_{xx}+3u_x w^2+6uww_x,\nonumber\\
70: w_t&=&-w_{xxx}+3w^2 w_x+3uw_x+3u_xw,
71: \eea
72: which arises as the classical part of one of superextensions of the
73: KdV equation. In this work, we study the integrability of this system
74: using the Painlev\'{e} test. Then, we use Dodd-Fordy [2] algorithm
75: of Wahlquist-Estabrook [3] prolongation technique in order to
76: obtain the Lax pair. We find a 3x3 matrix spectral problem
77: for the Kersten-Krasil'shchik system.\\ \\
78: %\baslik{2. Singularity analysis}
79: {\em Singularity Analysis}:
80: Let us study the integrability of (1)
81: following the Weiss-Kruskal algorithm of singularity analysis
82: [4], [5]. The algorithm is well known and widely used, therefore
83: we omit unessential computational details.
84:
85: First, we find that a hypersurface $\phi(x,t)=0$ is
86: non-characteristic for the system (1) if $\phi_x \neq 0$ and set
87: $\phi_x=1$ without loss of generality. Then we substitute the
88: expansions
89: \bea
90: u=u_0(t)\phi^{\alpha} + \ldots +
91: u_r(t)\phi^{r+\alpha} +\ldots, \nonumber\\
92: w=w_0(t)\phi^{\beta} + \ldots +
93: w_r(t)\phi^{r+\beta} +\ldots,
94: \eea
95: into (1), and find the
96: following branches (i.e. admissible choices of $\alpha, \beta, u_0$
97: and $w_0$), together with the positions $r$ of resonances (where
98: arbitrary functions can enter the expansions):
99: \bea
100: \alpha &=&-2,\; \; \beta=-1,\; \;u_0=1,\; \;w_0=\pm i, \nonumber\\
101: r &=& -1,1,2,3,4,6;
102: \eea
103: \bea
104: \alpha &=& -2,\;\; \beta=-1,\; \;u_0=2,\; \;w_0=\pm 2i, \nonumber\\
105: r &=& -2,-1,3,3,4,8;
106: \eea
107: \bea
108: \alpha &=& -2,\;\; \beta=2,\; \;u_0=2, \; \; \forall w_0(t), \nonumber\\
109: r &=& -4,-1,0,1,4,6;
110: \eea
111: \bea
112: \alpha &=& -2,\;\; \beta=3,\; \;u_0=2, \;\; \forall w_0(t), \nonumber\\
113: r &=& -5,-1,-1,0,4,6;
114: \eea
115: besides those which correspond to the Taylor expansions governed
116: by the Cauchy-Kovalevskaya theorem.
117:
118: The branch (3) is generic: the expansions (2) with (3) describe
119: the behavior of a generic solution near its singularity.
120: The non-generic branches (4), (5) and (6) correspond to singularities
121: of special solutions. The branches (4) and (5) admit the following
122: interpretation, in the spirit of [6]: (4) describes the collision of two
123: generic poles (3) with same sign of $w_0$, whereas (5) describes the
124: collision of two generic poles (3) with opposite signs of $w_0$. The
125: branch (6) corresponds to (5) with $w_0 \rightarrow 0$.
126:
127: Next, we find from (1) the recursion relations for the coefficients
128: $u_n(t)$ and $w_n(t)\; (n=0,1,2,\ldots)$ of the expansions (2),
129: separately for each of the branches, and check the consistency
130: of those recursion relations at the resonances. The recursion
131: relations turn out to be consistent, therefore the expansions
132: (2) of solutions of (1) are free from logarithmic terms. We
133: conclude that the system (1) passes the Painlev\'{e} test for
134: integrability successfully and must be expected to possess
135: a Lax pair.\\ \\
136: %\baslik{3. Prolongation structure}
137: {\em Prolongation Structure}:
138: By introducing the variables $p = u_x$, $q = w_x$, $r = p_x$, $s = q_x$,
139: we assume that there exist NxN matrix functions $F$ and $G$,
140: depending upon $u,w,p,q,r,s$, such that
141: \bea
142: y_x &=& - y F, \nonumber\\
143: y_t &=& - y G,
144: \eea
145: where $y$ is a row matrix with elements $y^A,
146: \; A =1,\ldots, N$.
147: The system of equations in (1) can be represented as the
148: compatibility conditions of (7) if
149: \be
150: F_t - G_x + [F,G]=0,
151: \ee
152: where $[F,G]$ is the matrix commutator. This
153: requirement gives the set of partial differential equations
154: for $F$ and $G$:
155: \bea
156: F_p = F_q = F_r = F_s = 0 ,\qquad
157: F_u = -G_r ,\qquad
158: 3wF_u+F_w =- G_s , \nonumber \\
159: pG_u + qG_w +rG_p + sG_q - 3(2up-qs+pw^2+2uwq)F_u \nonumber\\
160: -3(w^2q+uq+pw)F_w - [F,G] =0.
161: \eea
162: Next, we integrate equations (9) and find
163: \be
164: F=(uw-\frac{w^3}{2})X_1 +\frac{w^2}{2}X_2 + uX_3+ wX_4 + X_5,
165: \ee
166: where $X_1,X_2,X_3,X_4,X_5$ are constant matrices of integration.
167: It is immediately seen that $X_1$
168: is in the center of prolongation algebra [3]. Hence, we can take
169: it to be zero and find $G$ as,
170: \bea
171: G = (-r-ws-q^2+2u^2-w^4-w^2u)X_3 -(s-w^3-3uw)X_4 \nonumber\\
172: -(p+wq)X_6 - uw X_7 -(\frac{w^2}{2}+u) X_8 \nonumber\\
173: -q X_9 -\frac{w^2}{2} X_{10}- w X_{11} + X_0,
174: \eea
175: where $X_0$ is a constant matrix of integration.
176: The remaining elements are
177: \bea
178: X_6 = [X_5,X_3],\qquad
179: X_7 = [X_4,X_6], \qquad
180: X_8 = [X_5,X_6], \nonumber\\
181: X_9 = [X_5,X_4], \qquad
182: X_{10} = [X_4,X_9], \qquad
183: X_{11} = [X_5,X_9].
184: \eea
185: The integrability conditions impose the following restrictions on
186: $X_i$,\\$(i =0, \ldots , 11)$,
187: \bea
188: [X_2,X_3] = 0 , \quad
189: [X_5,X_0] = 0 , \quad
190: [X_3,[X_3,X_6]] = 0 , \quad
191: [X_2,[X_4,X_3]] = 0 , \nonumber\\
192: {}[X_3,[X_4,X_3]] = 0, \quad
193: [X_3,[X_4,[X_4,X_3]]]=0,\quad
194: [[X_4,[X_4,X_3]],[X_3,X_6]]=0,\nonumber\\
195: {}2X_6 +[X_5,X_2]=0,\quad
196: [X_3,X_0]-[X_5,X_8]=0,\quad
197: [X_4,X_2] + 4 [X_4,X_3] = 0 , \nonumber\\
198: {}[X_4,X_0]-[X_5,X_{11}] = 0 , \quad
199: 3X_6 - \frac{1}{2}[X_5,[X_3,X_6]]-[X_3,X_8] = 0,\nonumber\\
200: {}3X_2 - 3[X_4,[X_4,X_3]]-[X_2,X_6] + [X_3,X_6] = 0,\nonumber\\
201: {}X_7 + 2[X_5,[X_4,X_3]]-[X_3,X_9]=0, \nonumber\\
202: {}[X_2,X_0]-2[X_4,X_{11}]-[X_5,X_8]-[X_5,X_{10}] = 0,\nonumber\\
203: {}[X_2,[X_5,[X_4,X_3]]]+[X_2,X_7]+
204: \frac{1}{2}[X_2,[X_2,X_9]]=0,\nonumber\\
205: {}3X_9-[X_3,X_{11}]-[X_4,X_8]-[X_5,X_7]-
206: 2[X_5,[X_5,[X_4,X_3]]]=0,\nonumber\\
207: {}[X_3,X_7]+\frac{1}{2}[X_4,[X_3,X_6]]+
208: [X_3,[X_5,[X_4,X_3]]]=0,\nonumber\\ \nonumber\\
209: X_9-\frac{1}{2}([X_2,X_{11}]+[X_4,X_8]+
210: [X_4,X_{10}])-\nonumber\\
211: \frac{1}{3}([X_5,[X_5,[X_4,X_3]]]+
212: [X_5,X_7])-\frac{1}{6}[X_5,[X_2,X_9] ]=0,\nonumber\\ \nonumber\\
213: \frac{1}{2}[X_2,X_5]+\frac{1}{4}([X_2,X_8]+
214: [X_2,X_{10}])+\nonumber\\
215: \frac{1}{3}([X_4,X_7]+[X_4,[X_5,[X_4,X_3]]])+
216: \frac{1}{6}[X_4,[X_2,X_9]]=0,\nonumber \\ \nonumber\\
217: 3X_6-\frac{1}{2}([X_2,X_8]+[X_3,X_8]+[X_3,X_{10}])-
218: [X_4,X_7]-\nonumber\\
219: 2[X_5,[X_3,X_6]]-[X_4,[X_5,[X_4,X_3]]]-
220: 2[X_5,[X_4,[,X_4,X_3]]]=0,\nonumber \\ \nonumber\\
221: 8[X_4,X_3]+\frac{1}{4}[X_2,[X_2,X_9]]-
222: 2[X_4,[X_4,[X_4,X_3]]]-\nonumber \\
223: \frac{1}{6}([X_3,[X_2,X_9]]+11[X_4,[X_3,X_6]])=0.\nonumber\\
224: \eea
225: Together with the Jacobi identities we obtain further relations:
226: \bea
227: [X_2,X_6] +2 [X_3,X_6] = 0, \qquad
228: [X_4,X_{11}] - [X_5,X_{10}] = 0, \nonumber\\
229: {}[X_5,[X_3,X_6]] - [X_3,X_8] = 0,\qquad
230: [X_2,X_8] - [X_5,[X_2,X_6]]=0, \nonumber\\
231: {}[X_5,[X_4,X_3]]+[X_3,X_9] - X_7 = 0, \nonumber\\
232: {}-4[X_5,[X_4,X_3]] + [X_2,X_9] + 2X_7=0,\nonumber\\
233: {}[X_2,[X_5,[X_4,X_3]]]+2[[X_4,X_3],X_6]=0,\nonumber\\
234: {}[X_3,[X_5,[X_4,X_3]]]-[[X_4,X_3],X_6]=0,\nonumber\\
235: {}[X_3,[X_2,X_9]]-[X_2,[X_3,X_9]]=0,\nonumber\\
236: {}[X_4,X_3]=0, \qquad
237: [X_2,X_7]=0, \qquad
238: [X_3,X_7]=0, \nonumber\\
239: {}[X_3,X_{10}]=0, \qquad
240: [X_4,X_7]=0, \qquad
241: [X_5,X_7]=X_9,\nonumber\\
242: {}[X_2,[X_2,X_9]]=0,\qquad
243: [X_4,[X_3,X_6]]=0,\nonumber\\
244: {}[X_5,X_8]+[X_5,X_{10}]=0.
245: \eea
246: In order to find the Lie algebra generated by F and matrix
247: representations of the generators $\{X_i\}_0^{11}$, we follow
248: the strategy of Dodd-Fordy [3].
249: First we reduce the number of elements. By using equations (12)
250: -(14), we get $X_2 =-2X_3$. Next, we locate nilpotent and neutral
251: elements. The equations (12) and (13) together with $X_2 =-2X_3$
252: give that, $[X_5,X_3]=X_6$ and
253: $[X_3,X_6]=2X_3$, hence $X_3$ is nilpotent and $X_6$ is the
254: neutral element.
255: Let us note that the system of equations in (1) has the following
256: scale symmetry
257: \be
258: x \rightarrow \lambda^{-1} x , \qquad
259: t \rightarrow \lambda^{-3} t , \qquad
260: u \rightarrow \lambda^{2} u ,\qquad
261: w \rightarrow \lambda w , \nonumber
262: \ee
263: which implies that the elements $X_{i}$ must satisfy
264: \bea
265: X_0 \rightarrow \lambda^{3}X_0 , \qquad
266: X_3 \rightarrow \lambda^{-1} X_3 , \qquad
267: X_4 \rightarrow X_4 ,\qquad
268: X_5 \rightarrow \lambda X_5 ,\nonumber\\
269: {}X_6 \rightarrow X_6, \qquad
270: X_7 \rightarrow X_7 ,\qquad
271: X_8 \rightarrow \lambda X_8,\qquad
272: X_9 \rightarrow \lambda X_9,\nonumber\\
273: {}X_{10} \rightarrow \lambda X_{10},\qquad
274: X_{11} \rightarrow \lambda^2 X_{11},
275: \eea
276: where $\lambda$ is a constant.
277: By using the basis elements,
278: we try to embed the prolongation algebra into $sl(n+1,c)$. Starting
279: from the case $n=1$, we found that $sl(2,c)$ can not be
280: the whole algebra. The simplest non-trivial closure is in terms of
281: $sl(3,c)$. We take
282: \be
283: X_3 = e_{-\alpha_1}, \qquad
284: X_6 = h_1,
285: \ee
286: where we use the standart Cartan-Weyl basis [7] of $A_2$.
287: Together with the scale symmetries we find that
288: \bea
289: X_0 & = &-4c_2^2\lambda^4 e_{-\alpha_1}-36 c_1^3\lambda^3(h_1 + 2h_2)
290: -4c_2\lambda^2e_{\alpha_1},\nonumber\\
291: X_4 & = & d_1(h_1+2 h_2)+
292: d_2 \lambda^{-1}e_{\alpha_2}+
293: d_3\lambda^2 e_{-\alpha_1-\alpha_2},\nonumber\\
294: X_5 & = & e_{\alpha_1}+ c_1\lambda(h_1 + 2h_2)
295: +c_2\lambda^2e_{-\alpha_1},\nonumber\\
296: X_7 & = & d_2\lambda^{-1}e_{\alpha_2}+d_3\lambda^2
297: e_{-\alpha_1-\alpha_2},\\
298: X_8 & = &-2e_{\alpha_1}+2c_2\lambda^2e_{-\alpha_1},\nonumber\\
299: X_9 & = & d_2\lambda^{-1}e_{\alpha_1+\alpha_2}-d_3\lambda^2 e_{-\alpha_2}
300: +3c_1d_2 e_{\alpha_2}-3c_1d_3\lambda^3 e_{-\alpha_1-\alpha_2},
301: \nonumber\\
302: X_{10} & = & -d_2 d_3\lambda(h_1+2 h_2)-
303: 6c_1d_2d_3\lambda^2 e_{-\alpha_1}, \nonumber\\
304: X_{11} & = & (9c_1^2+c_2)d_2\lambda e_{\alpha_2} +6c_1d_3\lambda^3
305: e_{-\alpha_2}+ 6c_1d_2 e_{\alpha_1+\alpha_2}+
306: (9c_1^2+c_2)d_3\lambda^4 e_{-\alpha_1-\alpha_2},\nonumber
307: \eea
308: where $\{c_i\}_1^2$ and $\{d_i\}_1^3$ are constants with conditions
309: \be
310: d_1d_2=0, \qquad
311: d_1d_3=0, \qquad
312: d_2d_3=6c_1,\qquad
313: c_2=9c_1^2.
314: \ee
315: We choose $d_1=0$, $c_1=d_2=1$.
316: So that, $X_7=X_4$ and $X_0=-36\lambda^2X_5$. Then, we obtain
317: the matrix representations of the generators $X_{i}$ as
318: \bea
319: X_3 = \left( \begin{array}{ccc}
320: 0 & 0 & 0 \\
321: 0 & 0 & 0 \\
322: 1 & 0 & 0 \\
323: \end{array} \right), \qquad
324: X_4 = \left( \begin{array}{ccc}
325: 0 & 0 & 0 \\
326: -\lambda^{-1} & 0 & 0 \\
327: 0 & 6\lambda^2 & 0
328: \end{array} \right) , \nonumber \\
329: {}X_5 = \left( \begin{array}{ccc}
330: -\lambda & 0 & 1 \\
331: 0 & 2\lambda & 0 \\
332: 9\lambda^2 & 0 & -\lambda \\
333: \end{array} \right) , \qquad
334: X_6 = \left( \begin{array}{ccc}
335: 1 & 0 & 0 \\
336: 0 & 0 & 0 \\
337: 0 & 0 & -1 \\
338: \end{array} \right) , \nonumber \\
339: {}X_8 = \left( \begin{array}{ccc}
340: 0 & 0 & -2 \\
341: 0 & 0 & 0 \\
342: 18\lambda^2 & 0 & 0 \\
343: \end{array} \right), \qquad
344: X_9 = \left( \begin{array}{ccc}
345: 0 & 6\lambda^2 & 0 \\
346: -3 & 0 & \lambda^{-1} \\
347: 0 & -18\lambda^3 & 0 \\
348: \end{array} \right) , \nonumber \\
349: {}X_{10} = \left( \begin{array}{ccc}
350: 6\lambda & 0 & 0 \\
351: 0 & -12\lambda & 0 \\
352: -36\lambda^2 & 0 & 6\lambda \\
353: \end{array} \right),\qquad
354: X_{11} = \left( \begin{array}{ccc}
355: 0 & -36\lambda^3 & 0 \\
356: -18\lambda & 0 & 6 \\
357: 0 & 108\lambda^4 & 0 \\
358: \end{array} \right).
359: \eea
360: By substituting the matrix representations of the generators into
361: equations (10) and (11) we can construct the Lax pair,
362: $\Psi_x=X\Psi$,$\Psi_t=T\Psi$,
363: for the system (1), with the following matrices X and T:
364: \be
365: X=\left( \begin{array}{ccc}
366: \lambda & w \lambda^{-1} & w^2-u-9\lambda^2 \\
367: 0 & -2\lambda & -6w\lambda^2 \\
368: -1 & 0 & \lambda \\
369: \end{array} \right),
370: \ee
371: $T=\{\{p+wq+3\lambda w^2-36\lambda^3,
372: (w^3+2uw-s)\lambda^{-1}-3q-18\lambda w, r+ws+\\
373: q^2-2u^2+w^4+w^2u-9\lambda^2 w^2+
374: 18\lambda^2u+324\lambda^4\},
375: \{6q\lambda^2-36\lambda^3 w, -6\lambda w^2+\\72\lambda^3,
376: 6(s-w^3-2uw)\lambda^2-18q\lambda^3+108\lambda^4w\},
377: \{-w^2-2u+36\lambda^2, q\lambda^{-1}+6w,
378: -p-wq+3\lambda w^2-36\lambda^3\}\}$, where the matrix T
379: is written by rows and $X=-F^{\dagger}$,$T=-G^{\dagger}$,
380: $\Psi=y^{\dagger}$.
381:
382: The forms of X and T are unusual in the sense of the dependence
383: on $\lambda$. It is possible to obtain equivalent matrices by
384: the gauge transformation,
385: \be
386: X^{\prime}=SXS^{-1},\;\; T^{\prime}=STS^{-1},
387: \ee
388: where
389: \be
390: S=\left( \begin{array}{ccc}
391: 1 & 0 & 0 \\
392: 0 & 0 & -1 \\
393: 0 & \lambda^{-1} & 0 \\
394: \end{array} \right).
395: \ee
396: The result is
397: \be
398: X^{\prime}=\left( \begin{array}{ccc}
399: \lambda & u-w^2+9\lambda^2 & w \\
400: 1 & \lambda & 0 \\
401: 0 & 6\lambda w & -2\lambda \\
402: \end{array} \right),
403: \ee
404: $T^{\prime}=\{\{p+wq+3\lambda w^2-36\lambda^3,
405: -r-ws-q^2+2u^2-w^4-w^2u+9 \lambda^2 w^2-
406: 18\lambda^2u-324\lambda^4,
407: w^3+2uw-s-3q\lambda-18\lambda^2 w\},
408: \{w^2+2u-36\lambda^2,-p-wq+3\lambda w^2-36\lambda^3,
409: -q-6w\lambda\},\{6q\lambda-36\lambda^2 w,
410: -6(s-w^3-2uw)\lambda+18q\lambda^2-108\lambda^3 w,
411: -6\lambda w^2+72\lambda^3\}\}$.\\
412: %\baslik{4. Concluding remarks }
413: The matrix $X^{\prime}$ gives us exactly the spectral problem
414: for the KdV equation when $w=0$. But $X^{\prime}$
415: does not reduce to the one for mKdV equation when $u=0$.
416: This result should be expected because the
417: Kersten-Krasil'shchik system, when $u=0$, gives not only mKdV
418: equation, as stated in [1], but also an ordinary differential
419: equation in $w$. Finally, we note that the Lax pair obtained
420: from (7) with (24) is a true Lax pair since the parameter
421: $\lambda$ cannot be removed from $X^{\prime}$ by a gauge
422: transformation, as can be proven by a gauge-invariant
423: technique [8].\\
424: \baslik{Acknowledgements}
425: This work is supported in part by
426: the Scientific and Technical Research Council of Turkey (TUBITAK).
427:
428:
429: \begin{reference}
430: \item P. Kersten, J. Krasil'shchik, E-print(2000) arXiv:nlin.SI/0010041.
431: \item R. Dodd, A. Fordy, Proc.R.Soc.Lond., {\bf A 385},
432: (1983)389.
433: \item H.D. Wahlquist, F.B. Estabrook,
434: J.Math.Phys.,{\bf 16}, (1975)1.
435: \item J. Weiss, M. Tabor, G. Carnevale,
436: J.Math.Phys.,{\bf 24}, (1983)522.
437: \item M. Jimbo, M.D. Kruskal, T. Miwa,
438: Phys. Lett. {\bf A 92}, (1982)59.
439: \item A.C. Newell, M. Tabor, Y.B. Zeng,
440: Physica {\bf D 29},(1987)1.
441: \item J.E. Humphreys, {\bf "Introduction to Lie algebras
442: and representation theory"}, Springer-Verlag, NewYork, 1972.
443: \item S. Yu. Sakovich, J. Phys. {\bf A 28}, (1995)2861.
444:
445: \end{reference}
446:
447:
448:
449:
450: \end{document}
451:
452:
453:
454:
455:
456:
457: