astro-ph0306487/ms.tex
1: %%manuscript for preprint 2002/06/17  Version edited by Thomas
2: %\documentstyle[12pt,aaspp4,here]{article}
3: %\usepackage{here}
4: 
5: \documentclass[12pt,preprint]{aastex}
6: %\documentclass[manuscript]{aastex}
7: %\documentclass[preprint2]{aastex}
8: 
9: %\documentclass[12pt]{iopart}
10: \usepackage{epsfig}
11: 
12: %\begin{document}
13: 
14: %\begin {flushright}
15: %OU-TAP-70 \\
16: %December 1997
17: %\end{flushright}
18: 
19: \shorttitle{Universality in the distribution of caustics}
20: \shortauthors{Yano et al.}
21: \begin{document}
22: \title{Universality in the 
23: distribution of caustics \\in the expanding Universe}
24: 
25: \author{Taihei Yano$^1$, Hiroko Koyama$^1$, Thomas Buchert$^{1,2,3}$ 
26: and Naoteru Gouda$^1$}
27: 
28: \affil{$^{1}$
29: National Astronomical Observatory, 
30: Mitaka, Tokyo 181--8588, Japan\\
31: $^{2}$
32: Department of Physics and Research Center for the Early Universe, 
33: School of Science, The University of Tokyo, Bunkyo--ku, Tokyo 113--0033, Japan\\
34: $^{3}$
35: Theoretische Physik, Ludwig--Maximilians--Universit\"{a}t,\\
36: Theresienstrasse 37, D--80333 M\"unchen, Germany\\
37: E--mails: yano@pluto.mtk.nao.ac.jp, hiroko.koyama@nao.ac.jp,
38: buchert@theorie.physik.uni-muenchen.de, naoteru.gouda@nao.ac.jp}
39: 
40: \begin{abstract}
41: We numerically investigate the long--time evolution of density perturbations 
42: after the first appearance of caustics in an expanding cosmological model with
43: one--dimensional `single--wave' initial conditions.
44: Focussing on the time--intervals of caustic appearances
45: and the spatial distribution of caustics at subsequent times, we
46: find that the time--intervals of caustic appearances
47: approach a constant, i.e., their time--subsequent ratio converges to $1$;
48: it is also found that the spatial distribution of caustics at a given
49: time features some universality rules, e.g., the ratio between 
50: the position of the nearest caustic from the center and
51: that of the second nearest caustic from the center approaches a constant.
52: Furthermore we find some rules for the mass distribution for each caustic.
53: Using these universality constants we are in the position to predict the spatial 
54: distribution of caustics at an arbitrary time in order to give an estimate 
55: for the power spectral index in the fully--developed non--dissipative 
56: turbulent (`virialized') regime.
57: \end{abstract}
58: 
59: \noindent
60: {\it Subject headings}:
61: cosmology:theory -- large--scale structures -- self--similar -- non--dissipative
62: gravitational turbulence
63: 
64: \section{INTRODUCTION}
65: 
66: The evolution of large--scale structure in an expanding universe model is
67: one of the most important and interesting problems in cosmology.
68: It is generally believed that these structures have
69: formed owing to gravitational instability. After structure has formed,
70: the combined action of multi--stream forces and 
71: gravity is responsible for their stability in time 
72: (Buchert \& Dom\'\i nguez 1998), eventually leading to `virialized' structures
73: on certain spatial scales such as galaxies and clusters of galaxies.
74: Such equilibrium states, if they persist in time, are not necessarily 
75: in thermodynamical equilibrium, but are the consequence of a process that
76: was termed by Gurevich \& Zybin (see their review 1995 and refs. therein) 
77: {\it non--dissipative gravitational
78: turbulence}, emphasizing the role of multi--stream forces that counteract
79: the gravitational action and so eventually lead to equilibrium structures,
80: supported by velocity dispersion in collisionless systems.
81: 
82: Here we consider collisionless particles
83: as candidates for cold dark matter, and focus our interest on
84: universal behavior in the gravitational collapse process. 
85: The hope is to discover 
86: some rules that govern the multi--streaming hierarchy in a self--gravitating
87: system. In this research note we shall concentrate on the simplest model
88: of a plane--symmetric `single--wave' perturbation on an expanding 
89: universe model; the discovery of universality laws needs a detailed study at
90: high--spatial resolution. An advantage of plane symmetry is the 
91: existence of exact solutions that we shall incorporate in our numerical
92: scheme. Other aspects of multi--streaming have already 
93: been studied in three spatial dimensions (see, e.g., 
94: Kofman et al. 1994), and there is a plethora of works on powerlaw, 
95: fractal, or self--similar scaling properties of the density field, which are
96: all indirectly related to the understanding of multi--stream systems.
97: Also numerical simulations of self--gravitating systems 
98: practically evolve a multi--stream flow on smaller spatial scales.
99: 
100: Recently, Yano \& Gouda (1998) have studied the time evolution 
101: of density perturbations in the plane--symmetric system. They showed
102: that the power spectrum of the evolved perturbations can be
103: roughly separated into five regimes,
104: having initially power--law shape and a cutoff scale as shown in Figure 1.
105: 
106: 
107: According to these results, we can first separate the power spectrum roughly
108: into three regimes on the scales larger than the cutoff scale:
109: one is the linear regime ($k < k_{nl}$: Regime 1).
110: The value of the power index in this regime 
111: remains the initial power index, $n$.
112: The second regime is the single--caustic regime 
113: ($k_{nl} < k < k_{snl}$: Regime 2).
114: In this regime, the power index becomes $-1$ and is independent of the initial 
115: conditions. This result is caused by the appearance of caustics at this scale, 
116: and these caustics determine the power index of 
117: the power spectrum in this regime.
118: The third regime is the multi--caustics regime 
119: ($k_{snl} < k < k_{cut}$: Regime 3).
120: In this regime, caustics on various scales of waves determine
121: the power spectrum.
122: Therefore, on these scales, the power spectral index $\mu$ 
123: can be traced back to initial conditions.
124: 
125: On the scales smaller than the cutoff scale the problem is more involved. 
126: Nevertheless, we can roughly separate the power spectrum into 
127: two more regimes:
128: after the appearance of the first caustic, the power index in the regime 
129: of the scale smaller than the cutoff scale becomes $-1$.
130: This value is, as mentioned above, caused by the appearance of caustics.
131: However, thereafter, more and more caustics appear and so the separation of 
132: caustics becomes smaller and smaller.
133: Even after the appearance of many caustics, on scales smaller than the 
134: characteristic separation of the caustics ($k > k_{cs}$: Regime 5), 
135: the power index of the power spectrum can be obtained by the 
136: density profile around the singular point:
137: the power index on these scales also becomes $-1$, a result that can
138: be derived from catastrophe theory (or the Lagrange--singularity theory,
139: respectively), see, e.g., Arnol'd et al. 1982, Kotok \& Shandarin (1988), 
140: Gouda \& Nakamura (1988, 1989), Yano \& Gouda (1998). 
141: On the other hand, on the scales larger than the characteristic
142: separation of the caustics, in the fully developed `virialized' regime
143: ($k_{cut} < k < k_{cs}$: Regime 4),
144: the distribution of the singular points determines the power index on
145: these scales instead of the density profile around one singularity:
146: %Because in this regime, 
147: we may say that the smoothed density profile with the smoothing
148: scale ($k_{cut}<k<k_{cs}$) determines the power index.
149: For a comprehensive description of the above discussion, see
150: Yano \& Gouda (1998).
151: 
152: In this paper we focus on the fully developed `virialized' multi--stream 
153: regime ($k_{cut} < k < k_{cs}$)
154: to clarify the physical process of the appearances of the caustics,
155: and we are going to estimate a value for the power index 
156: of the power spectrum in this regime.
157: For this purpose we study the time--intervals of the appearances of the caustics 
158: as well as the spatial distribution of the caustics at subsequent times.
159: The initial condition is chosen as `single--wave' perturbation
160: defined by $v=v_0\sin x$,
161: where $v$ is the velocity, $x$ is the position of a sheet,
162: and $v_0$ is an arbitrary constant. The evolution of such a  perturbation 
163: determines the power spectral index in the `virialized' regime, 
164: essentially Region 4 in Figure 1.
165: Accordingly we succeed to find some universal rules about the time 
166: when each caustic appears,
167: the spatial distribution of the caustics at subsequent times, and the
168: distribution of mass for each caustic.
169: 
170: Our paper is organized as follows.
171: In $\S 2$, we will briefly explain the numerical method to calculate
172: the time evolution of density perturbations in the expanding 
173: universe model.
174: In $\S 3$ we show the numerical results providing some rules for the
175: time--appearances of caustics, the spatial distribution of caustics at
176: subsequent times, and the mass distribution for each caustic. Here,
177: we also evaluate the power index of the power spectrum in the `virialized'
178: regime. The final section is devoted to a summary and a brief outlook.
179: 
180: 
181: \section{MODEL AND NUMERICAL METHOD}
182: 
183: We investigate the time evolution of a plane--symmetric density perturbation in
184: an Einstein--de Sitter universe model by using a semi--numerical method.
185: In the one--dimensional system many plane--parallel sheets move only 
186: in perpendicular direction to the surface of these sheets.
187: When two sheets cross, they are allowed to pass through eachother freely.
188: In this sheet system, there is an exact solution
189: until two sheets cross over as follows
190: (Zel'dovich 1970, Sunyaev \& Zel'dovich 1972, Doroshkevich et.al. 1973, 
191: Zentsova \& Chernin 1980, Buchert 1989, Yano \& Gouda 1998):
192: \begin{eqnarray}  
193: x &=& q + B_1(t)S_1(q) + B_2(t)S_2(q), \nonumber \\
194: v &=& \dot{B_1}(t)S_1(q) + \dot{B_2}(t)S_2(q)\;\;, 
195: \label{1}
196: \end{eqnarray}  
197: where $q$ and $x$ are the Lagrangian and the (comoving) Eulerian coordinates, 
198: respectively.
199: Here, $S_1(q)$, and $S_2(q)$ are arbitrary functions of $q$. 
200: $B_1(t)$, and $B_2(t)$ are the growing mode and the decaying mode of linear 
201: perturbation solutions, respectively.
202: We are considering the Einstein--de Sitter background universe model, i.e., 
203: $B_1(t)=a$ and $B_2(t)=a^{-\frac{3}{2}}$,
204: where $a$ is the scale factor. The velocity $v$ is the peculiar--velocity
205: normalized by the scale--factor.
206: We can compute the crossing time of all neighboring pairs of sheets.
207: We use the shortest of these crossing times as a time step.
208: Then, we can compute the new positions and velocities for all sheets 
209: at this crossing time.
210: After two sheets cross, we exchange the velocities of two sheets that
211: just crossed.
212: Then we obtain again $S_1(q), S_2(q)$, and therefore exact solutions
213: as follows:
214: \begin{eqnarray}  
215: S_1(q)=\frac{3}{5}a^{-1}(x-q)+\frac{2}{5}\dot{a}^{-1}v, \nonumber \\
216: S_2(q)=\frac{2}{5}a^{\frac{3}{2}}(x-q)
217:  -\frac{2}{5}\dot{a}^{-1}a^{\frac{5}{2}}v\;\;.
218: \end{eqnarray}  
219: These new exact solutions can be used until again two sheets cross.
220: In this way we obtain the exact loci of the sheets 
221: by coupling these solutions.
222: This semi--numerical method has good accuracy, because 
223: we connect exact solutions.
224: 
225: In our calculation we are going to use $2^{12}$ sheets.
226: Periodic boundary conditions are fixed at a length of $2\pi$. 
227: Then we consider the evolution of the `single--wave' perturbation
228: whose wave number is $1$. 
229: 
230: \section{RESULTS}
231: 
232: In this section we calculate the time--intervals for the appearances of
233: caustics, the spatial distribution of caustics at subsequent times, and
234: the mass distribution for each caustic. Accordingly we find some rules
235: for the evolution of the caustic distribution. 
236: Using these rules, we are going to estimate the
237: power index of the power spectrum in the `virialized' regime.
238: 
239: \subsection{Universality rules and constants for the occurence of caustics}
240: 
241: We examine the time evolution of the `single--wave' perturbation
242: setting $S_1(q) =\sin q$ and $S_2(q) =0$ in Equations (\ref{1}).
243: Then, comoving positions $x$ and scaled peculiar--velocities $v$ of
244: each sheet are simplified to: 
245: \begin{eqnarray}  
246: x &=& q + B_1(t)\sin(q)\;\; , \nonumber \\
247: v &=& \dot{B_1}(t)\sin(q)\;\;,
248: \end{eqnarray}
249: see Figure 2-a.
250: Hereafter we call this case the 
251: {\it single--wave case}.
252: 
253: 
254: 
255: 
256: The first caustic has just appeared at the center of the x--axis ($x=0$)
257: in Figure 2-b. After the first appearance of a caustic, more caustics
258: appear at the center (phase mixing starts in phase space) (Figures 2-c -- 2-d)
259: (Doroshkevich et. al. 1980; Melott 1983; Gouda \& Nakamura 1989; Yano \& Gouda 
260: 1998). According to this cascade of appearing caustics, 
261: the density distribution 
262: evolves as in Figures (3-a -- 3-d).
263: Figures 2-a, 2-b, 2-c, and 2-d correspond to Figures 3-a, 3-b, 3-c, and 3-d, 
264: respectively.
265: As we see from Figures 3-a, 3-b, 3-c, and 3-d, a hierarchy of nested density
266: peaks appears in the course of time.
267: 
268: 
269: 
270: 
271: 
272: In order to discover rules for the distribution of the caustics at
273: an arbitrary time,
274: we sort the time--intervals in the appearances of the caustics, 
275: the spatial distribution of the caustics,
276: and the mass distribution for each caustic at a certain time 
277: as follows.
278: 
279: First, we examine the time--intervals related to the appearances of the caustics.
280: The solid line in Figure 4 shows the time when caustics appear
281: at the center of the x--axis (a so--called A3--type singularity).
282: The dashed line shows the time--intervals for the appearances of caustics.
283: As we can see from Figure 4, these time--intervals are almost constant.
284: In order to clarify this point, we introduce the ratio of those
285: intervals defined by $T_i \equiv (t_{i+2}-t_{i+1})/(t_{i+1}-t_i)$,
286: where $t_i$ is the time when the i--th caustics appears.
287: Then, the relation between $T_i$ and the time when the $i$--th caustic appears
288: is plotted in Figure 5.
289: We see a loss of accuracy after around the 10th caustic appeared.
290: Nevertheless, independent of the number of sheets,
291: we see that the value $T_i$ quickly converges to $1$.
292: We can understand why the ratio becomes 1 as follows:
293: according to the time evolution of the density distribution,
294: more and more caustics appear at the center of the x--axis and these
295: caustics may be viewed as forming a bound cluster.
296: After the evolution of this cluster decouples from the expansion of the 
297: universe model, sheets around the center of the cluster behave like a
298: harmonic oscillator due to the fact that an enhanced gravitational
299: potential forces particles with velocities less than the escape velocity
300: from the cluster to rebound. The periods 
301: of the oscillation of the sheets finally becomes independent of the amplitude.
302: Therefore, it is reasonable that the time--intervals of the appearances of
303: caustics becomes constant, i.e., the value of  $T_i$ becomes 1.
304: 
305: Second, we consider the spatial distribution of caustics.
306: In Figure 6 we plot the spatial positions of caustics when the 
307: N--th caustic appears at the center of the x--axis.
308: Here we note that the number of appearances of
309: caustics can be used instead of time, because we have already shown that the
310: intervals may be considered practically constant.
311: We define the position of a caustic, $X(i,j)$, as the $i$--th 
312: position of a caustic at the time when the $j$--th caustic appears. 
313: Then, for example, 
314: $X(i,i)$ represents the position of the $i$--th caustic 
315: that has just appeared as A3--type at the center of the x--axis.
316: Now, we examine the  spatial positions of caustics at a certain time ($N=N_0$).
317: We adopt $N_0=10$ in our analysis because the accuracy of the positions 
318: of caustics decreases for $N_0 > 10$ as shown in Figure 5.
319: The solid and dashed lines in Figure 7 denote the spatial
320: position of the $N$--th caustic, $X(N,10)$, and the spatial interval
321: between the (N-1)--th and the N--th caustics, $X(N-1,10)-X(N,10)$, respectively.
322: Furthermore we depict the corresponding ratios:
323: the solid and dashed lines in Figure 8 depict
324: the ratio of spatial positions $X(N,10)/X(N-1,10)$
325: and the ratio of spatial intervals
326: $(X(N-1,10)-X(N,10))/(X(N-2,10)-X(N-1,10))$, respectively.
327: In Figure 8 we see no constant in the spatial positions or 
328: spatial intervals of caustics at first glance.
329: 
330: However, we see a constant when we plot positions of the nearest
331: caustic from the center, $X(N-1,N)$, or the ratio $X(N-1,N)/X(N-2,N-1)$.
332: In Figure 9, the relation between 
333: the position $X(N-1,N)$ and the time when the N--th caustic appears is plotted.
334: The ratio of positions $X(N-1,N)/X(N-2,N-1)$ is plotted in Figure 10.
335: In Figure 10, $R_1$ represents the ratio between
336: the position of the nearest caustic from the center and
337: that of the second nearest caustic from the center $X(N-1,N)/X(N-2,N)$.
338: $R_2$ represents the ratio between the nearest caustic from the
339: center at the (N-1)th appearance of a caustic and that at the Nth appearance of
340: a caustic $X(N-1,N)/X(N-2,N-1)$.  Both ratios feature constants:
341: the ratio $R_1$ and $R_2$ have values of around $0.15$ and $0.3$, respectively.
342: 
343: If we plot positions of caustics by using a 
344: logarithmic scale of spatial positions and a linear scale of time, 
345: we can see some ``lattice'' as shown in Figure 11. This apparent 
346: regularity is the result of the existence of constants in the ratios.
347: Here we note that calculated positions are not exact for smaller
348: than about $10^{-6}$ because of finite sheet resolution.
349: Therefore, the ``lattice'' is destroyed after around the 10--th appearance of 
350: caustics.
351: 
352: %%%%%%%%%%%%%%%%%%%%%%
353: 
354: Third, we investigate the mass distribution for each caustic at a certain time
355: $(N=N_0)$.
356: Here, we define the mass, $M(i,N)$, as proportional to the number of sheets 
357: which are located inbetween $X(i,N)$ and $X(i+1,N)$ at the time, when the Nth 
358: caustic appears. The schematic
359: picture of $M$ is shown in Figure 12.
360: 
361: Then, we obtain the mass distribution for each caustic at $N_0=10$, which
362: is shown in Table \ref{tbl-1} and Figure 13.
363: We deduce the following relation between $i$ and $M$ for large values
364: of $i$:
365: \begin{equation}
366: M(i,N)\propto R_M^i, \qquad\qquad {\rm where} \qquad R_M\sim 0.5\;\;.
367: \end{equation}
368: Since we know the spatial distribution of the caustics and the mass
369: distribution for the caustics, we can construct the smoothed density profile 
370: to calculate the power index of the power spectrum in the `virialized' regime.
371: 
372: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
373: 
374: \subsection{Evaluation of the power spectrum in the `virialized' regime}
375: 
376: In this subsection we estimate the index of the power spectrum in the
377: `virialized' regime, using the rules which we have found in the
378: previous subsection.
379: Assuming that the spatial distribution of caustics obeys the law of 
380: a geometric series,
381: \begin{equation}
382: % x(n)=\left(0.15\right)^n
383:  X(i,N_0)=a_0a(N_0)^i \;\;,
384: \end{equation}
385: %near $x \to0$.
386: that is,
387: %$n \to x$
388: \begin{equation}
389: %n=\frac{\ln x}{\ln(0.15)}
390: i=\frac{\ln X(i,N_0)-\ln a_0}{\ln a}\;\;,
391: \end{equation}
392: and also that the mass distribution for each caustic obeys the law of 
393: another geometric series,
394: \begin{equation}
395: % m(n)=\left(0.5\right)^n
396:  M(i,N_0)=b_0b(N_0)^i\;\;,
397: \end{equation}
398: we find the density profile as:
399: \begin{eqnarray}
400: \label{density-profile}
401: \varrho(X)&=&\frac{M(i,N_0)}{X(i,N_0)-X(i+1,N_0)}%\nonumber\\
402: %&=&\frac{b^i}{a^i-a^{i+1}}\nonumber\\
403: %&=&\left(\frac{b}{a}\right)^i\frac{1}{1-a}\nonumber\\
404: =\frac{b_0}{(1-a)a_0}\left(\frac{b}{a}\right)^{-\ln a_0/\ln a}
405: X^{\frac{\ln (b/a)}{\ln a}}\;\;.
406: %\frac{\left(0.5\right)^n}
407: %{\left(0.15\right)^n-\left(0.15\right)^{n+1}}\nonumber\\
408: %&=&1/0.75 (1/0.3)^n\nonumber\\
409: %&=&x^{\frac{\ln 1/0.3}{\ln 0.15}}\nonumber\\
410: %&\sim& x^{-1.20397/1.89712}=x^{-0.634921}
411: \end{eqnarray}
412: Its Fourier transform gives 
413: %(Iwanami mathematical formula II p.275)
414: \begin{eqnarray}
415: \delta_k&\propto& \int \varrho(X)e^{ikX}dX\nonumber\\
416: &\propto& \int X^{\frac{\ln (b/a)}{\ln a}}e^{ikX}dX\nonumber\\
417: %&=&\sqrt{\frac{\pi}{2}}\sin \left(\frac{0.37\pi}{2}\right)\Gamma(1-0.37)
418: %k^{-0.63}
419: &\propto&k^{-1-\frac{\ln (b/a)}{\ln a}}\;\;.
420: \end{eqnarray}
421: Finally we obtain the power spectrum as:
422: %in the limit of $k\rightarrow k_{cut}$ is
423: \begin{eqnarray}
424: P(k)\equiv |\delta_k|^2 &\propto& k^{-2-2\frac{\ln (b/a)}{\ln a}}\;\;.
425: \end{eqnarray}
426: In fact, we have found  $a=R_1=0.15$ and $b=R_M=0.5$ from the previous 
427: subsection, so that we obtain the power spectrum in the explicit form: 
428: %in the virialized regime 
429: \begin{eqnarray}
430: \label{final}
431: P(k) \propto k^{\nu}, \qquad \qquad {\rm where} \qquad \nu=-0.73.
432: \end{eqnarray}
433: On scales smaller than the characteristic scale $k>k_{cs}$ (Regime 5 in
434: Figure 1), the power index of the power spectrum becomes $-1$. This
435: is because the density profile around one singular point determines 
436: the power spectrum on this scale.
437: 
438: On scales larger than the characteristic scale $k<k_{cs}$, 
439: on the other hand, the density profile in the spatial interval,
440: where the singular points lie closer than the scale of this interval, 
441: is smoothed, and the power index of the power spectrum
442: deviates from $-1$.
443: The larger scales we consider, the more  
444: the density profile is smoothed.
445: The power index in the `virialized' regime $k_{cut}<k<k_{cs}$,
446: accordingly, is determined by
447: the smoothed density profile (\ref{density-profile}), and therefore
448: the result for the power index of the power spectrum 
449: under the smoothed density profile is given by (\ref{final}).
450: 
451: Our result shows that the power index in the `virialized' regime 
452: $k_{cut}<k<k_{cs}$
453: may be approximated by the index value $\nu=-0.73$.
454: In addition, let us note that the wavenumbers $k_{cs}$, which represent the
455: characteristic separations of caustics, becomes larger and larger 
456: as time increases, since the characteristic separation of caustics
457: becomes shorter and shorter.
458: 
459: 
460: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
461: 
462: \section{SUMMARY AND OUTLOOK}
463: 
464: We have investigated the evolution of density perturbations
465: after the first appearance of a caustic, exemplified for a plane--symmetric
466: `single--wave' perturbation in an expanding universe model.
467: Following Yano \& Gouda (1998), who showed that the power spectrum can be
468: roughly separated into five regimes with different power--law index,
469: we have explored the fourth regime. 
470: The detailed analysis of a many--sheet system is needed in this regime in order
471: to determine the time--evolution and the asymptotics of the power spectral 
472: index.
473: 
474: We have argued that the system of caustics displays some universality,
475: after some time, with respect to regularity of time--intervals 
476: in the appearances of caustics, and the spatial distribution of caustics 
477: at subsequent times.
478: 
479: First, 
480: we have noticed the existence of a rule for the time--intervals 
481: of the appearances of caustics: 
482: the ratio of intervals quickly approaches $1$, a fact that we have made
483: plausible.  
484: Second, 
485: we have found some rules hidden in the spatial distribution of caustics:
486: the ratio between 
487: a position of the nearest caustic from the center and
488: that of the second nearest caustic from the center $R_1 =X(N-1,N)/X(N-2,N)$
489: approaches a value of around $0.15$.
490: Furthermore,
491: the ratio between the nearest caustic from the
492: center at the (N-1)th appearance of a caustic and that at the Nth appearance of
493: a caustic $R_2= X(N-1,N)/X(N-2,N-1)$ attains a value of around $0.3$.
494: 
495: Third, we have found a rule for the mass distribution for each caustic,
496: that is, we have derived the density distribution of a cluster.
497: The mass distribution for each caustic can be fit to a line with the geometric
498: progression of ratio $0.5$. 
499: 
500: Furthermore, using the above constants, 
501: we have found the smoothed density
502: profile, and finally have evaluated the power index of the power spectrum 
503: in the `virialized' regime to be $-0.73$.
504: This value is different from $-1$ that is determined by the density
505: profile around one singular point.
506: In other words, the power spectum in the region below
507: the cutoff scale can be divided into two clearly distinct regions, 
508: the `single--caustic' regime and the `virialized' regime.
509: These results are consistent with the numerical results 
510: by Yano \& Gouda (1998).
511: 
512: The discovery presented in this research note, namely  
513: that there exist remarkable rules for the distribution
514: of caustics, points towards several interesting theoretical implications
515: for long--term properties of self--gravitating systems.
516: As in studies of dynamical systems we here deal with a hierarchy of 
517: bifurcations. A fruitful way to explore universality properties further is
518: to use renormalization group techniques similar to those developed for 
519: the Feigenbaum scenario of pitchfork bifurcations (see: Feigenbaum 1978, 1979;
520: Derrida et al. 1979; Collet \& Eckmann 1980). 
521: With this note we have just provided a hint to the existence of universal
522: constants. A detailed renormalization group and fixpoint analysis is needed
523: to support these findings. In this context an interesting question has to 
524: remain open: the existence of chaotic regimes. 
525: 
526: Besides this theoretical interest we think that our results may be useful
527: to check the accuracy of numerical simulations concerning their 
528: stability under long--time integration.
529: 
530: In this paper we have focused on one non--trivial regime,
531: the `virialized' regime (Regime 4 in Figure 1), and have succeeded to evaluate 
532: the power index of the power spectrum in this regime semi--numerically.
533: In order to complete the understanding of the evolution of the power spectra 
534: of density fluctuations, we have also to understand the behavior of the power index
535: in the multi--wavenumbers regime (Regime 3 in Figure 1).
536: 
537: Furthermore, we could employ the systematic framework of the velocity 
538: moment hierarchy to eventually derive analytically universal behavior of
539: the density profile; also, nonlinear statistical methods beyond
540: the power spectrum would give more refined information.
541: 
542: \acknowledgments
543: 
544: {\small This work was supported in part by 
545: a Grant--in--Aid for Scientific Reserach (B) (No. 15340066).
546: TB acknowledges support by the National Observatory Tokyo where this work was
547: started, as well as hospitality at the University of Tokyo with
548: support by the Research Center for the Early Universe (RESCEU, Tokyo),
549: during which discussions on this work were continued. 
550: Also, partial support by the Deutsche Forschungsgemeinschaft 
551: (within SFB 375) is acknowledged.}
552: 
553: %\newpage
554: 
555: \begin{thebibliography}{2000}
556: 
557: \bibitem{} Arnol'd, V.I., Shandarin, S.F. \& Zel'dovich, Ya.B. 1982, 
558: Geophys. Astrophys. Fluid Dynamics 20, 111
559: 
560: \bibitem{} Buchert, T. 1989, Astron. Astrophys. 223, 9  
561: 
562: \bibitem{} Buchert, T. \& Dom\'\i nguez, A. 1998, Astron. Astrophys. 335, 395 
563: 
564: \bibitem{} Collet, P. \& Eckmann, J.P. 1980: Iterated Maps on the Interval as
565: Dynamical Systems, Progress in Physics 1, Birkh\"auser
566: 
567: \bibitem{} Derrida, B., Gervois, A., Pomeau, Y. 1979, J. Phys. A12, 269
568: 
569: \bibitem{}Doroshkevich, A.G., Ryaben'kii, V.S. \&  Shandarin, S.F. 1973, 
570: Astrophysics 9, 144
571: 
572: \bibitem{}Doroshkevich, A.G., Kotok, E.V., Novikov, I.D., Polyudov, A.N., 
573: Shandarin, S.F.\& Sigov, Yu.S. 1980, M.N.R.A.S. 192, 321
574: 
575: \bibitem{} Feigenbaum, M. 1978, J. Stat. Phys. 19, 25
576: 
577: \bibitem{} Feigenbaum, M. 1979, J. Stat. Phys. 21, 669
578: 
579: \bibitem{} Gouda, N. \& Nakamura, T. 1988, Prog. Theor. Phys. 79, 765
580: 
581: \bibitem{} Gouda, N. \& Nakamura, T. 1989, Prog. Theor. Phys. 81, 633  
582: 
583: \bibitem{} Gurevich, A.V. \& Zybin, K.P. 1995, Sov. Physics Uspekhi 38, 687
584: 
585: \bibitem{} Kofman L., Bertschinger, E., Gelb, J.M., Nusser, A. \& Dekel, A.
586: 1994, Ap.J. 420, 44
587: 
588: \bibitem{} Kotok, E.V. \& Shandarin, S.F. 1988, Soviet Astr. 32, 351 
589: 
590: \bibitem{} Melott, A.L. 1983, Ap.J. 264, 59
591: 
592: \bibitem{} Sunyaev, R.A. \& Zel'dovich, Ya.B. 1972, Astron. Astrophys. 20, 189
593: 
594: \bibitem{} Yano, T. \& Gouda, N. 1998, Ap.J.Suppl. 118, 267
595:   
596: \bibitem{} Zel'dovich, Ya.B. 1970, Astron. Astrophys. 5, 84
597: 
598: \bibitem{} Zentsova, A.S. \& Chernin, A.D. 1980, Astrophys. 16, 108
599: 
600: %Colombi, S., Bouchet, F.R. \& Hernquist,L. 1996, \apj, 465, 14 
601:   
602: %Couchman, H.M.P. \& Peebles, P.J.E. 1997, preprint (astro-ph/9708230)
603: 
604: %Davis, M., \& Peebles, P.J.E. 1977, \apjs, 34, 425 (DP)
605: 
606: %Fillmore, J.A., \& Goldreich, P. 1984, \apj, 281, 9 
607: 
608: %%Gouda, N. 1989, Prog.Theor.Phys., 81, 648  
609:   
610: %%Gouda, N.1996a, \apj, 469, 455   
611: 
612: %%Gouda, N.1996b, preprint (OU-TAP 47)
613: 
614: %Gouda, N. 1995, Prog.Theor.Phys., 94, 33
615:   
616: %Gouda, N. \& Nakamura, T. 1988, Prog.Theor.Phys., 79, 765  
617:   
618: %Gouda, N. \& Nakamura, T. 1989, Prog.Theor.Phys., 81, 633  
619: 
620: %Jain, B.,Mo, H.J. \& White, S.D.M., 1995, \mnras, 276, L25
621:   
622: %Kotok, E.V. \& Shandarin, S.F. 1987, Soviet Astr., 31, 600
623: 
624: %Kotok, E.V. \& Shandarin, S.F. 1988, Soviet Astr., 32, 351  
625: 
626: %Melott, A.L. 1983, \apj, 264, 59
627: 
628: %Peebles, P.J.E. 1980, The Large-Scale Structure of the Universe  
629: %(Princeton:Princeton Univ.Press)  
630: 
631: %Peebles, P.J.E. 1993, The Principles of Physical Cosmology  
632: %(Princeton:Princeton Univ.Press)  
633: 
634: %Roytvarf, A. 1987, Vestnik. Moscow University. Ser.1, No.1, 65
635: 
636: %Roytvarf, A. 1987, Vestnik. Moscow University. Ser.1, No.3, 41
637: 
638: %Shandarin, S.F. \& Melott, A.L. 1990, \apj 364, 396
639: 
640: %%Suto, Y. 1993, Prog.Theor.Phys., 90, 1173
641:    
642: %Yano, T. \& Gouda, N. 1997, \apj, 487, 473
643:   
644: %%Yess, C. \& Shandarin, S.F. 1996, \apj, 465, 2
645: 
646: %Zel'dovich, Ya.B. \& Shandarin, S.F. 1982 Sov. Astron. Lett., 8, 139
647: %\end{references}  
648: 
649: \end{thebibliography}
650: 
651: %\end{document}
652: %\newpage
653: \begin{figure}
654: \plotone{f1.eps}
655:  \caption{Schematic representation of the evolved power spectrum for a general 
656: initial power spectrum after the first appearance of caustics.\label{fig1}
657: }  
658: \end{figure}
659: 
660: \begin{figure}
661: \plottwo{f2a.eps}{f2b.eps}
662: 
663: \plotone{f2c.eps}
664: \plotone{f2d.eps}
665: %\plottwo{f2c.eps}{f2d.eps}
666: %\end{tabular}
667:  \caption{(a) Distribution in phase space for the `single--wave' case
668: at the initial time (a), at the first appearance of caustics (b),
669: at the second appearance of caustics (c), and 
670: at the fifth appearance of caustics (d).\label{fig2}
671: }  
672: \end{figure}
673: 
674: \begin{figure}
675: %\includegraphics[height=6cm]{f3a.eps}\includegraphics[height=6cm]{f3b.eps}
676: %\includegraphics[height=6cm]{f3c.eps}\includegraphics[height=6cm]{f3d.eps}
677: \plottwo{f3a.eps}{f3b.eps}
678: 
679: \plotone{f3c.eps}
680: \plotone{f3d.eps}
681:  \caption{ (a)Density distribution for the `single--wave' case 
682: at initial time (a), and at the first (b), the second (c) and the
683: fifth (d) appearance of caustics. \label{fig3}
684: }  
685: \end{figure}
686: 
687: 
688: 
689: \begin{figure}
690: %\includegraphics[height=13cm]{figure/fig4.eps}
691: \plotone{f4.eps}
692:  \caption{Times and intervals for the appearance of caustics at the center of 
693: the x--axis. The solid and dotted lines represent times and intervals, 
694: respectively. \label{fig4}
695: }  
696: \end{figure}
697: 
698: \begin{figure}
699: %\includegraphics[height=13cm]{figure/fig5.eps}
700: \plotone{f5.eps}
701:  \caption{Ratio of intervals in the appearance of caustics. 
702: The solid, long--dashed, short--dashed, dotted, and dash--dotted lines
703: are cases for different numbers of sheets: 
704: n=4095, n=2047, n=1023, n=511, and n=255, respectively.\label{fig5}
705: }  
706: \end{figure}
707: 
708: \begin{figure}
709: %\includegraphics[height=13cm]{figure/fig6.eps}
710: \plotone{f6.eps}
711:  \caption{Spatial distribution of caustics at the N--th appearance 
712: of caustics (the {\it bifurcation diagram}).\label{fig6}
713: }  
714: \end{figure}
715: 
716: \begin{figure}
717: %\includegraphics[height=13cm]{figure/fig7.eps}
718: \plotone{f7.eps}
719:  \caption{Spatial positions and intervals of caustics at the 10--th appearance 
720: of caustics. The solid and dashed lines represent the spatial positions
721: and intervals, respectively. Smaller N represents a position that is further
722: away from the center.\label{fig7}
723: }  
724: \end{figure}
725: 
726: \begin{figure}
727: %\includegraphics[height=13cm]{figure/fig8.eps}
728: \plotone{f8.eps}
729:  \caption{Ratio of spatial positions and intervals 
730: of caustics at the 10--th appearance 
731: of caustics. The solid and dashed lines denote ratios of spatial positions
732: and intervals, respectively.\label{fig8}
733: }  
734: \end{figure}
735: 
736: \begin{figure}
737: %\includegraphics[height=13cm]{figure/fig9.eps}
738: \plotone{f9.eps}
739:  \caption{Spatial position of the caustic that is nearest to 
740: the center of the x--axis $X(N-1,N)$.\label{fig9}
741: }  
742: \end{figure}
743: 
744: \begin{figure}
745: %\includegraphics[height=13cm]{figure/fig10.eps}
746: \plotone{f10.eps}
747:  \caption{Ratio of positions of caustics.
748: $R_1$ represents the ratio between
749: the position of the nearest caustic from the center and
750: that of the second nearest caustic from the center $X(N-1,N)/X(N-2,N)$.
751: $R_2$ represents the ratio between the nearest caustic from the
752: center at the (N-1)th and that at the Nth caustic appearance,
753: $X(N-1,N)/X(N-2,N-1)$.\label{fig10}
754: }  
755: \end{figure}
756: 
757: \begin{figure}
758: %\includegraphics[height=13cm]{figure/fig11.eps}
759: \plotone{f11.eps}
760:  \caption{Spatial positions of caustics at the N--th appearance of caustics.\label{fig11}
761: }  
762: \end{figure}
763: 
764: \begin{figure}
765: %\includegraphics[height=10cm]{figure/fig12.eps}
766: \plotone{f12.eps}
767: \caption{Schematic picture of the density distribution and 
768: the definition of the mass for the i--th caustic (the shaded region).\label{fig12}}  
769: \end{figure}
770: 
771: 
772: %\begin{table}
773: % \begin{tabular}{|c|c|c|}
774: %order of caustics&position of caustics&numbers of sheets\\
775: %\hline
776: %1& 0.2736580 &1094\\
777: %2& 0.1288644 &2021\\
778: %3&0.0507813&1888\\
779: %4&0.0170624&1289\\
780: %5&0.0051794&763\\
781: %6&0.0014272&414\\
782: %7&0.0003467&204\\
783: %8&0.0000683&95\\
784: %9&0.0000081&42\\
785: %\hline
786: % \end{tabular}
787: %\caption{Mass distribution for each caustic when the 10th caustics appear.}
788: %\end{table}
789: \begin{figure}
790: %\includegraphics[height=13cm]{figure/fig13.eps}
791: \plotone{f13.eps}
792:  \caption{Mass distribution contributing to each caustic at the 10--th
793:  appearance of caustics.\label{fig13}}  
794: \end{figure}
795: 
796: \begin{table}
797: \begin{center}  
798: \caption{Mass distribution for each caustic when the 10th caustics appear.}  
799: \label{tbl-1}
800: \begin{tabular}{ccc}
801: \hline
802: \hline
803: order of caustics&position of caustics&numbers of sheets\\
804: \hline
805: 1& 0.2736580 &1094\\
806: 2& 0.1288644 &2021\\
807: 3&0.0507813&1888\\
808: 4&0.0170624&1289\\
809: 5&0.0051794&763\\
810: 6&0.0014272&414\\
811: 7&0.0003467&204\\
812: 8&0.0000683&95\\
813: 9&0.0000081&42\\
814: \hline
815: \end{tabular}
816: \end{center}
817: \end{table}
818: 
819: \end{document}  
820: 
821: