1: \documentclass[12pt, preprint]{aastex}
2:
3: \begin{document}
4:
5: \title{Virialization in Dark Energy Cosmology}
6:
7: \author{Peng Wang}
8:
9: \affil{\it Department of Physics, Stanford University, Stanford, CA
10: 94305-4060, USA} \email{pengwang@stanford.edu}
11:
12: \begin{abstract}
13: We discuss the issue of energy nonconservation in the virialzation
14: process of spherical collapse model with homogeneous dark energy. We
15: propose an approximation scheme to find the virialization radius. By
16: comparing various schemes and estimating the parameter
17: characterizing the ratio of dark energy to dark matter at the
18: turn-around time, we conclude that the problem of energy
19: nonconservation may have sizable effects in fitting models to
20: observations.
21: \end{abstract}
22:
23: \keywords{cosmology:theory-galaxies:clusters:general-large-scale
24: structure of universe-galaxies:formation }
25:
26: \maketitle
27:
28: \begin{center}
29: {\bf I. INTRODUCTION}
30: \end{center}
31:
32: Analyzing the effects of dark energy on the nonlinear structure
33: formation process may provide us new ways of constraining the
34: properties of dark energy. Especially, a lot of recent works focused
35: on analyzing the effects of dark energy in the framework of the
36: spherical collapse model (\citealt{LLPR, Steinhardt, Maor, Mota,
37: Horellou, Battye, Iliev, Kamion, Nunes}). The spherical collapse
38: model is a simple but powerful framework to understand the growth of
39: bound systems in the universe \citep{Gunn}. It is also incorporated
40: in the famous Press-Schecter formalism \citep{PS}.
41:
42: In spherical collapse model, we consider a top-hat spherical
43: overdensity with mass $M$ and radius $R$. At early times, it expands
44: along with the Hubble flow and density perturbations grow
45: proportionally to the scale factor. After the perturbation exceeds a
46: critical value, the spherical overdensity region will decouple from
47: the Hubble flow and go through three phases: (1) expansion to a
48: maximum radius, $R_{ta}$, after which the overdensity will
49: turn-around to collapse; (2) collapse; (3) virialization at the
50: virialization radius $R_{vir}$.
51:
52: A key parameter of spherical collapse model is the ratio of the
53: virialized radius and turn-around radius $x=R_{vir}/R_{ta}$. Let's
54: first review briefly the derivation of the standard result $x=0.5$
55: in Einstein-de Sitter cosmology \citep{Peacock}.
56:
57: As is well-known, the self-energy of a sphere of nonrelativistic
58: particles with mass $M$ and radius $R$ is
59: \begin{equation}
60: V_{mm}=-{3\over5}{GM^2\over R}\label{0}
61: \end{equation}
62:
63: After the system virializes, the virial theorem $T_{vir}=(R/2)dU/dR$
64: and Eq.~(\ref{0}) tells us that $T_{vir}=-1/2U_{vir}$. Substituting
65: this to the energy conservation equation $T_{vir}+U_{vir}=U_{ta}$,
66: we get the standard result $x=0.5$.
67:
68:
69: When considering the evolution of spherical overdensities in the
70: presence of homogeneous dark energy (since \citet{Caldwell} have
71: shown that generally dark energy does not cluster on scales less
72: than 100 Mpc), the gravitational potential energy of the spherical
73: dark matter overdensity will be modified by a new term due to the
74: gravitational effects of dark energy on dark matter \citep{Maor,
75: Mota}:
76: \begin{equation}
77: U_{mQ}={1\over2}\int\rho_m\Phi_QdV \label{pe}
78: \end{equation}
79: where $\Phi_{Q}$ is the potential induced by dark energy
80: \begin{equation}
81: \Phi_Q=-{2\pi G}(1+3\omega_Q)\rho_Q \left(R^2-{r^2\over3}\right).
82: \label{poten}
83: \end{equation}
84:
85: Note that if we do not consider the pressure of the dark energy, as
86: \citet{Steinhardt} did, then potential is proportional to $\rho$,
87: i.e. the factor $1+3\omega_Q$ should be unity in front of the
88: expression for $\Phi_{Q}$. In a fully relativistic treatment,
89: pressure will also contribute to gravitation, so the potential is
90: proportional to $\rho+3p$ \citep{Maor, Mota}.
91:
92: Thus the total potential energy of spherical overdensity is
93: \begin{equation}
94: U=U_{mm}+U_{mQ}=-{3\over5}{GM^2\over R}-(1+3\omega_Q){4\pi
95: G\over5}M\rho_Q R^2,\label{totalpe}
96: \end{equation}
97: where we have substituted Eq.~(\ref{poten}) into Eq.~(\ref{pe}).
98:
99: In most of the current literature \citep{Horellou, Battye, Iliev},
100: in the presence of smooth dark energy, $x$ is still found by using
101: the energy conservation equation
102: \begin{equation}
103: U_{vir}+T_{vir}=U_{ta}.\label{}
104: \end{equation}
105:
106:
107: Then from the virialization theorem
108: $T_{vir}={R_{vir}\over2}{dU(R)\over dR}\mid_{R=R_{vir}}$ and
109: Eq.~(\ref{totalpe}), we can find
110: \begin{equation}
111: T_{vir}=-{1\over2}U_{mm,vir}+U_{mQ,vir}.\label{T}
112: \end{equation}
113:
114: Substituting Eq.~(\ref{T}) into the energy conservation equation,
115: one can find:
116: \begin{equation}
117: {1\over2}U_{mm,vir}+2U_{mQ,vir}=U_{mm,ta}+U_{mQ,ta},\label{}
118: \end{equation}
119: from which we can find the equation determining $x$
120: \begin{equation}
121: -4q(1+3\omega_Q)y^{-3(1+\omega_Q)}x^3+2[1+(1+3\omega_Q)q]x-1=0\label{old}
122: \end{equation}
123: where we have defined $q=\rho_{Q,ta}/\rho_{mc,ta}$ and $y=a_{vir}/
124: a_{ta}$. If we set the $(1+3\omega_Q)$ factor to unity in the first
125: two terms of Eq.~(\ref{old}), we can get the equation found by
126: \citet{LLPR, Steinhardt}.
127:
128: \begin{center}
129: {\bf II. THE ENERGY NON-CONSERVATION PROBLEM}
130: \end{center}
131:
132: The procedure described in Sec. I is problematic when dark energy is
133: dynamical, i.e. $\omega_Q\neq-1$. Indeed, when $\omega_Q>-1$,
134: $\rho_Q$ is decreasing with time. When considering collapse of dark
135: matter halo of the cluster scale, since dark energy does not cluster
136: below 100 Mpc \citep{Caldwell}, $\rho_Q$ in $U_{mQ}$ should take its
137: background value, i.e. evolving with time. In other words, dark
138: energy does not virialize with dark matter, otherwise it cannot be
139: smooth. Thus, $U_{mQ}$ will contribute a non-conservative force to
140: the dark matter particle. So the clustering dark matter with
141: potential (\ref{totalpe}) is a non-conservative system:
142: \begin{equation}
143: U_{vir}+T_{vir}< U_{ta}.\label{}
144: \end{equation}
145:
146: So actually, in the presence of dark energy, dark matter cannot
147: reach virialization in the strict sense. But for dark matter halo of
148: the cluster scale, it clusters at the era when the effect of dark
149: energy is still small. So it is reasonable to assume that dark
150: matter particles can reach a quasi-equilibrium state in which virial
151: theorem holds instantaneously. This is supported by the observations
152: of relaxed cluster in our Universe (see e.g. \citet{allen}). In the
153: following discussion, we still call this quasi-equilibrium state as
154: virialization. Thus assuming dark matter has reached this
155: quasi-equilibrium state, its total energy can be computed by the
156: virial theorem,
157: \begin{equation}
158: U=U_{vir}+{R\over2}{dU\over R}\mid_{R=R_{vir}}=-{3\over10}{GM^2\over
159: R_{vir}}-(1+3\omega_Q){4\pi
160: G\over5}MR_{vir}^2\rho_{Q,ta}\left(a(t)\over
161: a_{ta}\right)^{-3(1+\omega_Q)},\label{qq}
162: \end{equation}
163: which is decreasing with time. Although this non-conservation effect
164: is small in our discussion, it is worth commenting that in dark
165: energy dominated era, this effect may be large. For example, in the
166: extreme case of phantom dark energy models, the effect of dark
167: energy may be so large that cluster, galaxy and even our solar
168: system will de-virialize in the future \citep{kamion2}.
169:
170: So using Eq.~(\ref{old}) to find $x$ will generally overestimate its
171: actual value. In fact, from the dark matter potential
172: (\ref{totalpe}), we can see that $U_{total,
173: vir}=U_{mm,vir}/2+2U_{mQ,vir}$ is a monotonically increasing
174: function of $x$. Thus, while in fact we have $U_{total,
175: vir}<U_{total, ta}$, if we still use $U_{total,vir}=U_{total,ta}$ to
176: determine $x$, we will get a $x$ larger than its actual value.
177:
178: If the dark energy density is time-independent, then the system with
179: the potential (\ref{totalpe}) is conservative and thus we can use
180: energy conservation legitimately. Thus to estimate the virialization
181: radius when dark energy density is changing with time, we can take
182: $\rho_{Q,vir}$ to be the same as $\rho_{Q,ta}$.
183:
184: With this approximation, the equation determining $x$ is,
185: \begin{equation}
186: -4q(1+3\omega_Q)x^3+2[1+(1+3\omega_Q)q]x-1=0.\label{main}
187: \end{equation}
188: It is worth commenting that taking $\rho_{Q,vir}$ to be
189: $\rho_{Q,ta}$ does not mean ignoring the background evolution of
190: dark energy, i.e. make it degenerate with a true cosmological
191: constant. First, there is a factor $1+3\omega_Q$ in Eq.~(\ref{main})
192: which is different from the case of a cosmological constant. Second,
193: the value of $\rho_{Q,ta}$ is different from $\rho_{Q,0}$; while for
194: a true cosmological constant, $\rho_{\Lambda}$ is constant all the
195: times. Thus using Eq.~(\ref{main}) to estimate $x$ can give us a
196: more realistic value than Eq.~(\ref{old}) and at the same is able to
197: discriminate among different values of $\omega$.
198:
199: \begin{figure}
200: % Requires \usepackage{graphicx}
201: \epsscale{.80}
202: \plotone{f1.eps}
203: \caption{Ratio of the virialization radius to the turn-around radius $x=R_{vir}/R_{ta}$
204: as a function of $q$, which characterizes the strength of dark energy at turn-around, for
205: $\omega_Q=-0.8$. The dotted line is computed by Eq.~(\ref{old}), the solid one is computed by Eq.
206: (\ref{main}), the dashed line is computed by Eq.~(28) of ~\citet{Maor} and the dashed-dotted line is computed by Eq.~(\ref{1.1})}\label{1}
207: \end{figure}
208:
209: Figure 1 shows the virialization radius to the turn-around radius
210: $x=R_{vir}/R_{ta}$ as a function of $q$. From the figure, it can be
211: seen that when $q$ is large, i.e. the effects of dark energy is
212: large, Eq.~(\ref{old}) (dotted line) will always predict a larger
213: virialization radius than Eq.~(\ref{main}) (solid line). This showed
214: explicitly the comment following Eq.~(\ref{qq}). Thus by assuming
215: the dark energy density to be constant during the virialization
216: process, we can use energy conservation and we can find a lower $x$
217: which is closer to the actual one.
218:
219: It is also interesting to note that all the curve in Fig.~1 is above
220: the standard value $x=0.5$ in an Einstein-de Sitter universe. This
221: is easy to understand. Since dark energy will cause an effective
222: repulsive force on the dark matter, the dark matter particles can
223: reach equilibrium with a larger radius. Thus, $x>0.5$ is a smoking
224: gun of dark energy (see also \citet{Maor}, which reaches similar
225: conclusion).
226:
227: Recently, \citet{Maor} considered the possibility that even if dark
228: energy does not fully cluster, it still fully virialize. Then there
229: will also be a energy non-conservation problem because the
230: virialized system (now containing both dark matter and dark energy)
231: does not cluster in the same rate. Note that this energy
232: non-conservation problem is different from the energy
233: non-conservation problem discussed above. In \citet{Maor}'s
234: analysis, since both dark matter and dark energy virialize, they
235: included the dark energy self-energy when using the virial theorem.
236: In our case, where dark energy is smooth and do not virialize, the
237: physical picture of what's going on is just a spherical overdensity
238: of dark matter particles collapsing in the background of smooth dark
239: energy. From the derivation of the virial theorem \citep{Marion}, we
240: know that the potential energy appearing in the virial theorem is
241: the one that will give rise to forces on the virialized particles.
242: So if dark energy does not virialize, we should consider only the
243: potential energies giving rise to dark matter self-gravitation and
244: its gravitational interaction with dark energy. This is why our
245: energy corrected equation (\ref{main}) is very different from that
246: of \citet{Maor} (Eq.~(28) in it).
247:
248: In Fig. 1, we showed the prediction of the virialization equation
249: found by \citet{Maor} (dashed line). It can seen that for $q$ not
250: too large, the prediction of \citet{Maor} is much smaller than our
251: result Eq.~(\ref{main}). This is conceivable. Since in
252: \citet{Maor}'s analysis, the positive self-energy of dark energy is
253: also included in the total potential energy of virialized particles,
254: $U_{total}$ will be larger than its actual value. Since $U_{total}$
255: is a monotonically increasing function of $x$, with the same initial
256: energy, the equation of \citet{Maor} will thus predict a smaller
257: $x$.
258:
259: \citet{Mota} have considered the case of fully clustered and
260: virialized dark energy. We think it is interesting to consider the
261: case that only a portion of dark energy cluster and virialize. In
262: this case, we should include the dark energy self-energy that will
263: cluster since this part of the self-energy will contribute to the
264: force felt by the virialized dark energy particles.
265:
266: In this case, the dark energy evolution equation is
267: \begin{equation}
268: \dot\rho_Q+(1-F)3{\dot R\over R}(1+\omega_Q)\rho_Q+F3{\dot a\over
269: a}(1+\omega_Q)\rho_Q=0,\label{}
270: \end{equation}
271: where $F$ characterizes the fraction of dark energy that cluster.
272: This equation can be integrated to find
273: \begin{equation}
274: \rho_Q=\rho_{Q,ta}x^{-3(1+\omega)(1-F)}y^{-3(1+\omega)F}\label{}
275: \end{equation}
276:
277: First, we use our approximation method: we neglect the background
278: evolution of the dark energy, i.e., we take
279: \begin{equation}
280: \rho_Q=\rho_{Q,ta}x^{-3(1+\omega)(1-F)}y^{-3(1+\omega)F}\rightarrow
281: \rho_{Q,ta}x^{-3(1+\omega)(1-F)},\label{}
282: \end{equation}
283: in the virialization process. Taking into account the observation
284: that we should only include a fraction $1-F$ of the dark energy
285: potential energy, we can get the equation determining $x$
286: \begin{eqnarray}
287: &&-(1-F)(1+3\omega_Q)\left[7-6(1+\omega_Q)(1-F)\right]q^2x^{-6\omega_Q+6F(1+\omega_Q)}\cr
288: &&-(2+3\omega_Q-F)[4-3(1+\omega_Q)(1-F)]qx^{-3\omega_Q+3F(1+\omega_Q)}\cr
289: &&+2[1+(1+3\omega_Q)q+(1-F)q+(1-F)(1+3\omega_Q)q^2]x-1=0\label{gamma}
290: \end{eqnarray}
291:
292: For $F=1$, Eq.~(\ref{gamma}) will reduce to Eq.~(\ref{main}), while
293: for $F=0$, it will reduce to the equation found by \citet{Mota}.
294: Thus, our result Eq.~(\ref{main}) can be continuously connected to
295: the case that dark energy will also collapse with dark matter. This
296: is physically satisfying.
297:
298: Second, if we adopt the proposal of restoring energy conservation by
299: \citet{Maor} then when $F=1$, the virialization equation is
300: \begin{equation}
301: -2(1+3\omega_Q)qx^{-3\omega_Q}-2(1+3\omega_Q)qy^{-3(1+\omega_Q)}x^3+[1+(1+3\omega_Q)q]x-1=0\label{1.1}
302: \end{equation}
303:
304: We showed the dependence of $x$ on $q$ from Eq.~(\ref{1.1}) as the
305: dashed-dotted line in Fig. 1. It can be seen that although the
306: approach of restoring energy conservation are different in Eqs.
307: (\ref{1.1}) and (\ref{main}), their predictions are rather close.
308: This illustrates that although the underlying ideas are different,
309: in practice, our approximation scheme is quantitatively close to the
310: scheme of \citet{Maor}. In both cases, the difference from the old
311: result (\ref{old}) is large when $q$ is large.
312:
313: From Fig. 1 we can also see that for $q\sim10^{-2}$ or smaller, we
314: get $x=0.5$ in all the four approaches. This is reasonable. In the
315: virialization process, it is the self-energy of matter that plays
316: the dominant role. In fact, $U_{QQ.vir}/U_{mQ,vir}$ and
317: $T_{QQ,vir}/T_{mQ,vir}$ are both of the order
318: \begin{equation}
319: (1+3\omega_Q)q\left({a_{vir}\over
320: a_{ta}}\right)^{-3(1+\omega_Q)}\left({R_{vir}\over
321: R_{ta}}\right)^3\label{}
322: \end{equation}
323: since $\left({a_{vir}\over a_{ta}}\right)^{-3(1+\omega_Q)}\simeq
324: 1.6^{-3(1+\omega_Q)}$ and $\left({R_{vir}\over R_{ta}}\right)^3\sim
325: 0.1$, for $q<0.01$ the above ratio is much smaller than $1$, and
326: thus we can expect that for small $q$, the problem of energy
327: conservation will not influence the virialization process greatly.
328:
329: \begin{figure}
330: % Requires \usepackage{graphicx}
331: \epsscale{.80}
332: \plotone{f2.eps}
333: \caption{The dependence of $q$ on $z_{vir}$ from Eq.~(\ref{q}). The solid, dashed, dotted lines correspond to
334: $\omega_Q=-0.7,-0.8,-0.9$, respectively.}\label{1}
335: \end{figure}
336:
337: Thus to estimate the effects of dark energy on virialization, and
338: especially the ambiguity of energy-nonconservation, it is necessary
339: to estimate the value of $q$ for the virialization redshift
340: $z_{vir}$ that would be interesting to observations. If for
341: observationally interesting $z_{vir}$, $q$ will always be quite
342: small, then we can conclude that the problem of energy
343: non-conservation will not bother us too much in analyzing the
344: effects of dark energy on the formation of non-linear structure.
345: Unfortunately, this is not the case.
346:
347: Let's begin by writing $q$ as
348: \begin{equation}
349: q={\rho_{Q,ta}\over\rho_{mc,ta}}={\rho_{Q,ta}\over
350: \zeta\rho_{m,ta}}={\Omega_{Q0}(1+z_{ta})^{3\omega_Q}\over\zeta\Omega_{m0}},\label{q}
351: \end{equation}
352: where we have defined $\zeta=\rho_{mc,ta}/\rho_{m,ta}$.
353:
354: First, after specified $z_{vir}$, $z_{ta}$ can be computed using the
355: fact that $t_{vir}=2t_{ta}$, which is due to the observation that
356: collapse proceeds symmetrically to the expansion phase. Then, we use
357: the fitting formula for $\zeta$ presented by \citet{Steinhardt}:
358: \begin{equation}
359: \zeta=\left({3\pi\over4}\right)^2\Omega_{m,ta}^{-0.79+0.26\Omega_{m,ta}-0.06\omega_Q}.\label{}
360: \end{equation}
361:
362: With those two inputs, we can get the dependence of $q$ on $z_{vir}$
363: from Eq.~(\ref{q}) shown in Fig. 2. It can be seen that $q$ will be
364: of the order $10^{-2}$ when the virialization redshift is larger
365: than roughly $2$. Combining this with Fig. 1, we conclude that for
366: observationally interesting clusters, i.e. clusters formed after
367: redshift $2$, the presence of dark energy will have sizable
368: modifications to the standard result $x=0.5$. Thus observational
369: evidence for $x>0.5$ would be a strong evidence in favor of
370: dynamical dark energy.
371:
372: \begin{figure}
373: % Requires \usepackage{graphicx}
374: \epsscale{.80}
375: \plotone{f3.eps}
376: \caption{Thick lines: the dependence of $\Delta_{vir}$ on $z_{vir}$ for
377: $\omega_Q=-0.6,-0.8, -1$ from top to bottom using Eq.~(\ref{main});
378: thin lines: the dependence of $\Delta_{vir}$ on $z_{vir}$ for
379: $\omega_Q=-0.6,-0.8$ from top to bottom using Eq.~(\ref{old}), the
380: case of $\omega=-1$ is identical with that of using
381: Eq.~(\ref{main}). }\label{1}
382: \end{figure}
383:
384: An important parameter in fitting theoretical calculations to
385: observation is the density contrast at virialization
386: $\Delta_{vir}\equiv\rho_{mc,vir}/\rho_{m,vir}=\zeta y^3x^{-3}$. Fig.
387: 3 shows the dependence of $\Delta_{vir}$ on $z_{vir}$ for
388: $\omega_Q=-0.6,-0.8, -1$ from top to bottom using Eq.~(\ref{main})
389: and Eq.~(\ref{old}). We can see that for $\omega_Q$ not too close to
390: $-1$, there is obvious differences in the predicted $\Delta_{vir}$
391: using Eq.~(\ref{main}) and Eq.~(\ref{old}). Thus, whether including
392: the effect of energy conservation may have large impact on fitting
393: models to cosmological observation such as weak lensing
394: \citep{Kamion}. It is worth commenting that if we ignore the
395: $1+3\omega$ factor in $U_{mQ}$, as in \citet{Steinhardt, Kamion},
396: the difference will be even larger. For example, \citet{Kamion}
397: computed that for $\omega_Q=-0.6$, $\Delta_{vir}\sim 420$ for
398: $z_{vir}=0$. Furthermore, as can be seen in Fig. 3, whatever scheme
399: we use to find $x$, there is notable difference between the case of
400: a true cosmological constant and dynamical dark energy. Thus,
401: cluster observations may provide important information on the
402: dynamical behavior of dark energy.
403:
404: \begin{center}
405: {\bf III. CONCLUSIONS AND DISCUSSIONS}
406: \end{center}
407:
408: To summarize, in this work we discussed the issue of energy
409: non-conservation in the virialzation process of spherical
410: overdensity with homogeneous dark energy. We proposed that taking
411: the dark energy density to be constant during the virialization
412: process to obtain an estimate of the virialization radius. By
413: comparing various schemes and estimating the parameter $q$, we
414: conclude that there will be sizable effect of dark energy on
415: virialization process. A general signature of dark energy is that
416: the final virialization radius will be larger than half of the
417: turn-around radius.
418:
419: It should be emphasized that the analysis in this work is quite
420: qualitative. More detailed numerical simulations and analysis of
421: observational data are required to estimate quantitatively the
422: effect of dark energy on spherical collapse models and answer the
423: general question ``can we constrain the evolution of dark energy by
424: studying the structures of non-linear objects in our Universe".
425: Furthermore, establishing firmly the result $x>0.5$ from observation
426: is challenging. In practice, it is much easier to measure directly
427: baryons in clusters. But there are some astrophysical processes
428: leading to energy non-conservation in the virialization process of
429: baryon in the dark matter halo (e.g. X-ray emission of the hot gas,
430: conduction, AGN heating, dynamical friction, etc., see e.g.
431: \citet{Kamion2} and references therein). We should compare the
432: effects of those processes to dark energy in realistic analysis. To
433: achieve this, we need to know the concrete physical mechanism of
434: virialization in both the dark matter and baryon sectors, which is
435: now still not well-understood. Thus more works in this direction is
436: needed and will be rewarding.
437:
438: \begin{center}
439: {\bf ACKNOWLEDGEMENT}
440: \end{center}
441: I would like to thank Irit Maor for helpful comment on the first
442: version of this paper.
443:
444: \begin{thebibliography}{99}
445:
446: \bibitem[Battye \& Weller (2003)]{Battye}
447: Battye, R. A., Weller, J., 2003, Phys.~Rev.~D, 68, 083506.
448:
449: \bibitem[Caldwell et al. (2003)]{kamion2}
450: Caldwell, R. R., Kamionkowski, M., Weinberg, N. N., 2003,
451: Phys.Rev.Lett. 91, 071301.
452:
453: \bibitem[Dave et al. (2002)]{Caldwell}
454: Dave, R., Caldwell, R. R., Steinhardt, P. J., 2002, Phys.~Rev.~D,
455: 66, 023516.
456:
457: \bibitem[Fabian \& Allen (2003)]{allen}
458: Fabian, A. C., Allen, S. W., astro-ph/0304020.
459:
460: \bibitem[Gunn \& Gott (1972)]{Gunn}
461: Gunn, J. E., Gott, J. R. I., 1972, \apj, 176, 1.
462:
463: \bibitem[Horellou \& Berge (2005)]{Horellou}
464: Horellou, C., Berge, J., astro-ph/0504465.
465:
466: \bibitem[Iliev \& Shapiro (2001)]{Iliev}
467: Iliev, I. T., Shapiro, P. R., 2001, \mnras, 325, 468.
468:
469: \bibitem[Lahav et al. (1991)]{LLPR}
470: Lahav, O., Lilje, P. B., Primack, J. R., Rees, M. J., 1991, \mnras,
471: 251, 128.
472:
473: \bibitem[Marion (1970)]{Marion}
474: Marion, J. B., 1970, \emph{Classical Dynamics of Particles and
475: Systems}, Second Edition, Academic Press, Sec. 7.15.
476:
477: \bibitem[Maor \& Lahav (2005)]{Maor}
478: Maor, I., Lahav, O., astro-ph/0505308.
479:
480: \bibitem[Mota \& de Bruck (2004)]{Mota}
481: Mota, D. F., van de Bruck, C., 2004, \aap, 421, 71
482: [astro-ph/0401504].
483:
484: \bibitem[Nunes \& Mota (2004)]{Nunes}
485: Nunes, N. J., Mota, D. F., astro-ph/0409481.
486:
487: \bibitem[Peacock (1999a)]{Peacock}
488: Peacock, J. A., 1999a, \emph{Cosmological Physics}, Cambridge
489: University Press, Sec. 15.8.
490:
491: \bibitem[Peacock (1999b)] {Kamion2}
492: Peacock, J. A., 1999b, \emph{Cosmological Physics}, Cambridge
493: University Press, Sec.~17.3.
494:
495: \bibitem[Press \& Schechter (1974)]{PS}
496: Press, W. H., Schechter, P., 1974, \apj, 187, 425.
497:
498: \bibitem[Wang \& Steinhardt (1998)]{Steinhardt}
499: Wang, L., Steinhardt, P. J., 1998, \apj, 508, 483.
500:
501: \bibitem[Weinberg \& Kamionkowski (2003)]{Kamion}
502: Weinberg, N. N., Kamionkowski, M., 2003, \mnras, 341, 251.
503:
504: \end{thebibliography}
505:
506:
507:
508:
509:
510: \end{document}
511: