1: %%
2: %% Beginning of file 'sample.tex'
3: %%
4: %% Modified 03 Jan 01
5: %%
6: %% This is a sample manuscript marked up using the
7: %% AASTeX v5.x LaTeX 2e macros.
8:
9: %% The first piece of markup in an AASTeX v5.x document
10: %% is the \documentclass command. LaTeX will ignore
11: %% any data that comes before this command.
12:
13: %% The command below calls the preprint style
14: %% which will produce a one-column, single-spaced document.
15: %% Examples of commands for other substyles follow. Use
16: %% whichever is most appropriate for your purposes.
17:
18: \documentclass[12pt,preprint]{aastex}
19:
20: %% manuscript produces a one-column, double-spaced document:
21:
22: % \documentclass[manuscript]{aastex}
23:
24: %% preprint2 produces a double-column, single-spaced document:
25:
26: % \documentclass[preprint2]{aastex}
27:
28: %% If you want to create your own macros, you can do so
29: %% using \newcommand. Your macros should appear before
30: %% the \begin{document} command.
31: %%
32: %% If you are submitting to a journal that translates manuscripts
33: %% into SGML, you need to follow certain guidelines when preparing
34: %% your macros. See the AASTeX v5.x Author Guide
35: %% for information.
36:
37: \newcommand{\vdag}{(v)^\dagger}
38: \newcommand{\myemail}{lia@paxi.cnrs-mrs.fr}
39:
40: %% You can insert a short comment on the title page using the command below.
41:
42: %\slugcomment{Not to appear in Nonlearned J., 45.}
43:
44: %% If you wish, you may supply running head information, although
45: %% this information may be modified by the editorial offices.
46: %% The left head contains a list of authors,
47: %% usually a maximum of three (otherwise use et al.). The right
48: %% head is a modified title of up to roughly 44 characters. Running heads
49: %% will not print in the manuscript style.
50:
51: \shorttitle{Bar-Halo Interaction and Bar Growth}
52: \shortauthors{Athanassoula}
53:
54: %% This is the end of the preamble. Indicate the beginning of the
55: %% paper itself with \begin{document}.
56:
57: \begin{document}
58:
59: %% LaTeX will automatically break titles if they run longer than
60: %% one line. However, you may use \\ to force a line break if
61: %% you desire.
62:
63: \title{Bar-Halo Interaction and Bar Growth}
64:
65: %% Use \author, \affil, and the \and command to format
66: %% author and affiliation information.
67: %% Note that \email has replaced the old \authoremail command
68: %% from AASTeX v4.0. You can use \email to mark an email address
69: %% anywhere in the paper, not just in the front matter.
70: %% As in the title, you can use \\ to force line breaks.
71:
72: \author{E. Athanassoula}
73: \affil{Observatoire de Marseille, 2 Place Le Verrier, 13248 Marseille
74: cedex 04, France}
75:
76: %% Notice that each of these authors has alternate affiliations, which
77: %% are identified by the \altaffilmark after each name. Specify alternate
78: %% affiliation information with \altaffiltext, with one command per each
79: %% affiliation.
80:
81: %% Mark off your abstract in the ``abstract'' environment. In the manuscript
82: %% style, abstract will output a Received/Accepted line after the
83: %% title and affiliation information. No date will appear since the author
84: %% does not have this information. The dates will be filled in by the
85: %% editorial office after submission.
86:
87: \begin{abstract}
88: I show that strong bars can grow in galactic discs, even when the
89: latter are immersed in haloes whose mass
90: within the disc radius is comparable to, or larger than, the mass of the
91: disc. I argue that this is due to the response of
92: the halo and in particular to the destabilising influence of the halo
93: resonant stars. Via this instability mechanism the halo can stimulate,
94: rather than restrain, the growth of the bar.
95: \end{abstract}
96:
97: %% Keywords should appear after the \end{abstract} command. The uncommented
98: %% example has been keyed in ApJ style. See the instructions to authors
99: %% for the journal to which you are submitting your paper to determine
100: %% what keyword punctuation is appropriate.
101:
102: \keywords{galaxies: evolution, galaxies: halos, galaxies: kinematics and dynamics}
103:
104: %% From the front matter, we move on to the body of the paper.
105: %% In the first two sections, notice the use of the natbib \citep
106: %% and \citet commands to identify citations. The citations are
107: %% tied to the reference list via symbolic KEYs. The KEY corresponds
108: %% to the KEY in the \bibitem in the reference list below. We have
109: %% chosen the first three characters of the first author's name plus
110: %% the last two numeral of the year of publication as our KEY for
111: %% each reference.
112:
113:
114: \section{Introduction}
115:
116: Galactic discs are generally unstable and form bars within a few
117: dynamical timescales (e.g. Athanassoula 1984, or Sellwood \&
118: Wilkinson 1993, for reviews). The first remedy proposed against this
119: instability is to immerse the disc in a massive halo component.
120: Ostriker \& Peebles (1973), from an insightful
121: analysis of $N$-body simulations with a very small number of
122: particles, were the first to outline the
123: stabilising influence of haloes. They further suggested that halo-to-disc
124: mass ratios of 1 to 2.5 (interior to the disc) are required for stability.
125: Athanassoula \& Sellwood (1986) measured the growth rates of the most
126: unstable bar-forming modes in 2D $N$-body simulations and
127: showed that the growth rate is
128: smaller in cases with large halo-to-disc mass ratios. Unfortunately,
129: no such analysis has been made for fully self-consistent 3D $N$-body
130: simulations.
131: On the other hand, Athanassoula \& Misiriotis (2002, hereafter AM) and
132: Athanassoula (2002a,b) argue that stronger bars can grow
133: in simulations with initially very high halo-to-disc mass ratios than
134: in simulations with lower halo-to-disc mass ratios.
135: In order to understand whether the above pieces of evidence are
136: contradictory, and, if so, to attain a coherent picture,
137: I will examine in this letter the role of the halo in the formation of
138: the bar. Several papers have focused on the influence of the
139: halo on the slowing down of the bar (e.g. Tremaine \& Weinberg 1984,
140: Weinberg 1985, Combes et al. 1990, Little \& Carlberg 1991, Hernquist
141: \& Weinberg 1992, Athanassoula 1996, Debattista \& Sellwood 1998 and
142: 2000). This paper is complementary, since it concentrates on the role
143: of the halo in determining the bar growth.
144: The discussion will focus on three simulations. The results, however,
145: are based on a much larger set of simulations, leading to the same
146: conclusions.
147:
148: \section{Results on bar formation}
149: \label{sec:simul}
150:
151: I will describe the results of three fully self-consistent 3D $N$-body
152: simulations
153: of isolated galaxies consisting of a disc and a halo component. The initial
154: conditions and the numerical methods are described in AM,
155: whose notation I follow here. The
156: three initial conditions have identical exponential
157: discs of unit mass and scale length, scale height $z_0$ = 0.2 and $Q$ =
158: 1. Their halos are five
159: times as massive as their discs and have different degrees of central
160: concentration. They are initially isotropic, spherical and
161: non-rotating, with a radial density profile
162:
163: \begin{equation}
164: \rho_h (r) = \frac {M_h}{2\pi^{3/2}}~~ \frac{\alpha}{r_c} ~~\frac {exp(-r^2/r_c^2)}{r^2+\gamma^2}.
165: \end{equation}
166:
167: \noindent
168: with $M_h$ = 5, $r_c$ = 10. The parameter $\alpha$ is a normalisation
169: constant defined by
170:
171: \begin{equation}
172: \alpha = [1 - \sqrt \pi~~exp (q^2)~~(1 -erf (q))]^{-1},
173: \end{equation}
174:
175: \noindent
176: where $q=\gamma / r_c$ and $erf$ is the error function (Hernquist
177: 1993). The initial circular
178: velocity curves of the three models are shown in the upper
179: panels of Fig.~1. The first simulation, hereafter MD, has $\gamma$ =
180: 5, and its disc dominates in the inner parts, roughly up to $r$ = 5. In
181: the other two (hereafter
182: MH and RH, respectively), $\gamma$ = 0.5, and the halo contribution is
183: slightly larger
184: than that of the disc up to the maximum of the disc rotation curve, and
185: considerably so at larger radii.
186: Model RH is identical to model MH,
187: except that its halo is rigid -- i.e. given by a potential imposed
188: on the disc particles -- and does not evolve during the simulation.
189: The number of particles in the simulations is 1131206, 1163030
190: and 200000, respectively. The simulations were run on a GRAPE-5 system
191: with a softening of 0.0625 and a time-step of 0.015625. I assessed
192: the numerical robustness of my results by trying
193: double the number of particles, different values of the softening
194: and time step, as well as direct summation and a non-GRAPE tree code.
195: For a reasonable calibration (AM),
196: $t$ = 500 corresponds to 7 $\times 10^9$ years.
197: Different values, however, can be obtained for different scalings of
198: the disc mass and scale length.
199:
200: The second row of panels of Fig.~1 gives the three circular
201: velocity curves at time $t$ = 800. Note that the disc material has moved
202: considerably inwards in the two first cases, but very little
203: in the third one. In model MD the difference between times 0 and 800 is
204: quantitative, since the disc dominates the inner parts both initially and
205: after the evolution, albeit to a larger degree after the bar has grown. On
206: the other hand for model MH the difference
207: is qualitative, since at $t$ = 800 the inner regions are dominated by
208: the disc, contrary to what was the case initially. The third row of
209: panels gives the face-on view of the disc component. The differences
210: between the properties of the three models are striking. A
211: comparison of models MD and MH shows that {\it the bar that grew
212: in the initially more halo-dominated environment is stronger
213: than the bar that
214: grew in the disc-dominated environment}. It is longer and its
215: isophotes are more rectangular-like. The strength of the bar can be
216: measured with the help of the relative amplitude of the
217: Fourier components of the density or mass (AM), and the last row of
218: panels in Fig.~1 shows that indeed the MH bar is considerably stronger
219: than the MD one. The $m$ = 4, 6 and 8
220: Fourier components of the face-on density of model MH are
221: also considerably stronger than those of model MD (AM).
222:
223: Very strong differences are also found when comparing models MH
224: and RH. Model MH has a very strong bar, while model RH has a
225: very mild oval in the innermost parts. Their edge-on views are also very
226: different (not shown here, cf. Athanassoula 2002b for another
227: example). Model MH seen side-on has a strong peanut or `X'-shape, and
228: a big bulge-like protuberance if seen end-on. Model RH shows no such
229: features. Finally the difference between their integrated Fourier
230: components is striking.
231:
232: The three models differ also in the way their bars evolve. In model MD the bar
233: grows very rapidly during the first part of the evolution, but little,
234: if at all, after that. On the contrary the growth of the bar in model MH
235: during the first part of the evolution is slower than in MD, but stays
236: considerable, although less important than in the first part, till the
237: end of the simulation. The pattern
238: speed of model MH starts off higher than that of model MD, but ends
239: smaller. The pattern speed of MD also decreases with
240: time, but less so (cf. Debattista \& Sellwood 1998, 2000). The pattern
241: speed of the non-axisymmetric component in
242: RH can not be measured reliably before $t$ = 300, and after
243: that does not show any signs of decrease. For models MD and MH
244: there is exchange of energy and
245: angular momentum between the disc and halo components, so that the
246: halo, which was initially non-rotating, displays rotation after the
247: bar has grown. This is small for model MD and considerable for model
248: MH, for which at $t$ = 900 the halo has somewhat less than
249: a half of the angular momentum of the disc, i.e. not far from a
250: third of the total.
251:
252: \section{The role of the halo}
253: \label{sec:halo}
254:
255: In order to understand the role of the halo in the formation and
256: evolution of the bar, I froze the
257: potential at four selected times during each simulation and chose randomly
258: 100\,000 disc and 100\,000 halo particles. I followed their orbits during
259: 40 bar rotation periods and
260: calculated their basic frequencies, namely
261: the angular frequency $\Omega$, the epicyclic frequency $\kappa$ and
262: the vertical frequency $\kappa_z$. For the two latter ones I used a
263: spectral analysis technique (Binney \& Spergel 1982, Laskar
264: 1990). In most cases there were several secondary peaks, and the
265: frequency was determined by the main one. The angular
266: frequency proved more difficult to calculate
267: reliably, so I supplemented
268: the spectral analysis with other, more straightforward methods, based
269: on following the angle as a function of time. Agreement between
270: the values for the angular frequency obtained by the various methods
271: was found to be satisfactory for 85\% or more of the particles.
272:
273: An orbit is resonant if there are three integers
274: $l$, $m$ and $n$ such that
275: $l \kappa + m \Omega + n \kappa_z = m \Omega_p$, where $\Omega_p$ is
276: the pattern speed of the bar. I will here
277: restrict myself to radial (planar) resonances, for which $n$ = 0. The most
278: important such resonances are the inner
279: Lindblad resonance (ILR), where $l$ = -1 and $m$ = 2, corotation
280: resonance (CR), where $l$ = 0, and the outer Lindblad resonance (OLR),
281: where $l$ = 1 and $m$ = 2. At CR the
282: particles have the same angular frequency as the bar, while at the
283: other resonances they make $m$ radial oscillations in the time they
284: make $l$ revolutions around the center of the galaxy.
285: Fig.~2 shows the number of particles (orbits),
286: $N_R$, that have a frequency ratio $R = (\Omega - \Omega_p) / \kappa$
287: within a bin of a given width centered on a value of this ratio,
288: plotted as a function of $R$.
289:
290: Let me first describe the results for the disc components of MD and MH.
291: The distribution in both cases is far from homogeneous, with
292: strong peaks at the location of the main resonances. The highest peak,
293: both for the MH and the MD disc, is for $(\Omega - \Omega_p) / \kappa$
294: = 0.5, i.e. at the ILR. Indeed, orbits
295: making two radial oscillations in the time they make one revolution
296: around the center of the galaxy are the backbone of the bar. In
297: principle $x_2$ type orbits (cf. Contopoulos \& Grosb{\o}l 1989 for a review)
298: could also be found in this peak. We have, however, verified that the
299: vast majority of the orbits here are $x_1$ type. The ILR peak
300: in the MH case is roughly 1.5 times higher than in MD, consistent with
301: the fact that the bar in the MH case is stronger.
302: Model MD has a sizeable CR peak and a much
303: lower one at the OLR, while model MH has only a small CR peak and no
304: OLR one. The ratio of the height of the CR peak to that
305: of the ILR peak is roughly 0.6 for the MD case, while for model MH it is only
306: 0.09. These differences are, to a large extent,
307: due to the differences in the corotation radii of the two models.
308: Thus at time 800 the CR radius is 7.1 for model
309: MH and 4.6 for model MD. Similar differences are found for the
310: OLR radii of the two models. Therefore these two resonances are in the
311: outer parts of the MH disc and can trap only few particles. This
312: is not the case for model MD, and the differences in the trappings are
313: reflected in the differences in the heights of the respective resonant peaks.
314: Model MH has also a clear peak at the inner 1:3 resonance and
315: perhaps one at the 1:4. Model MH has a clear inner 1:4 peak.
316:
317: The big surprise, however, comes from the halo component. So far
318: considered as non-, or little, responsive, it shows, on the contrary,
319: unmistakable signs of strong resonances with the bar. For both models
320: the strongest resonance is CR, which proves to be much stronger in the
321: halo than in the disc component. In fact the CR peak in the halo
322: of model MH is higher than any of the disc peaks in MD. There
323: are also sizeable peaks at the OLR.
324:
325: There are important differences between the orbital structures of the MH and MD
326: haloes, as was the case for the corresponding discs. The ILR peak of
327: model MH is relatively high -- the second highest peak for this model
328: and time -- and, in
329: general, there is considerably more material between
330: ILR and CR than in model MD, while the CR peak is 40\% higher than in
331: model MD. All these are in agreement with the fact
332: that the MH halo is much more concentrated than the MD one, while
333: the MH corotation is at a larger radius. For smaller values of $R$
334: the situation shifts and the peaks are stronger for the MD
335: case. The OLR peak of model MD is more than 1.8 times
336: as high as the MH one, and the -1:1 peak is also clear in MD,
337: while for model MH this resonance and its surroundings have been
338: depleted.
339:
340: The relative and absolute heights of the resonance peaks
341: vary with time. However, a large fraction of the orbits which are
342: within a resonance peak at a given time continue to be so at a later
343: time. Thus, comparing times 500 and 560, I find that 60 to 80 per cent of
344: all orbits in one of the main resonant peaks at time 500 are found in the
345: same peak at time 560. Similar numbers can be found by comparing times
346: 800 and 860, except for the CR peak of the halo in the MH
347: simulation, where the fraction falls to 55 per cent. Seen the
348: uncertainties in delimiting the extent of a frequency peak, and the
349: uncertainties in the calculation of the frequencies, these numbers are
350: compatible with a large trapping of particles in the resonances. A
351: more detailed description of these changes, together
352: with a discussion on their effect on the structure and the dynamical
353: evolution of the galaxy will be given elsewhere.
354:
355: \section{Discussion}
356: \label{sec:discuss}
357:
358: In the above I compared three simulations starting off with
359: identical discs, but different halo components. Any differences in
360: their dynamical evolution should thus be attributed to the haloes.
361: The strongest bar forms in the most halo-dominated case, provided this
362: is live, while the simulation with an identical but rigid halo forms
363: only a mild oval distortion, in the inner parts only. This very big
364: difference can be attributed only to the responsiveness of the halo.
365: The disc-dominated case, i.e. the case with the less important halo
366: component, formed an intermediate bar. I thus reach the interesting
367: conclusion that haloes can, at least in some cases, stimulate the bar
368: instability and lead to stronger bars.
369:
370: Lynden-Bell \& Kalnajs (1972) have shown that disc stars at resonances
371: can absorb or emit angular momentum, thus driving the dynamical
372: evolution. Halo stars in resonance can also exchange
373: energy and angular momentum (Tremaine \& Weinberg 1984). In general,
374: the halo resonances will absorb angular momentum. Prompted by such
375: considerations I searched for resonant stars in the halo, and found
376: large numbers (at least after the bar has grown). Since these can
377: exchange energy
378: and angular momentum with disc resonant stars, they can stimulate the
379: bar instability, contrary to previous beliefs, and thus explain why
380: stronger bars can grow in more halo-dominated surroundings. The
381: situation is particularly clear in the case of model MH, where, as
382: Fig.~2 shows, there are hardly any absorbers in the disc to take the
383: angular momentum emitted by the stars at the ILR. This task is thus
384: necessarily performed by the halo resonant stars.
385:
386: Since the bar is inside corotation, it has negative energy and angular
387: momentum (e.g. Lynden-Bell \& Kalnajs 1972), and if it emits angular
388: momentum it will be destabilised, i.e. it will in general grow stronger. This
389: is in good agreement with the results on angular momentum transfer
390: discussed at the end of \S2. These show that the halo takes
391: positive angular momentum from the disc/bar component and, since the
392: disc, the bar and the final halo component all rotate in the same direction,
393: this will destabilise the bar.
394:
395: In the initial phases of the evolution the bar grows faster
396: in the disc dominated surroundings, in good agreement with previous
397: results (e.g. Athanassoula \& Sellwood 1986). However, in fine, the
398: bar in the MH model reaches a higher amplitude, due to the stronger
399: exchange of energy and angular momentum with the resonant halo stars.
400: Thus a massive halo may help the bar grow and
401: become very strong, so that very strong bars may be found in initially
402: halo-dominated galaxies. In other words there is no disagreement
403: between the results of e.g. Athanassoula \& Sellwood (1986) and those
404: presented here or in AM or in Athanassoula (2002a, b).
405:
406: Hernquist \& Weinberg (1992) checked the angular momentum given to a
407: live halo by a rigid bar turned first on and then off adiabatically. In
408: spite of considerable noise, they found indications that the angular
409: momentum was deposited primarily at the resonances and that it is
410: mainly absorbed. This tentative result is clearly established
411: here, where both the disc and the halo are live, the
412: evolution is not artificially forced and the resonances are clearly
413: outlined.
414:
415: In the linear theory the exchange of energy and angular momentum is
416: linked with the growth of the wave. In the nonlinear theory on
417: the other hand it has been linked to the slowdown of the bar
418: (Tremaine \& Weinberg 1984, Weinberg 1985, Hernquist \& Weinberg 1992).
419: Both effects are clearly visible in the simulations presented here and
420: should be physically linked.
421:
422: The orbital structures revealed in Fig.~2 are not specific to
423: the models discussed above. I have repeated a similar analysis for
424: other times and other simulations and found similar behaviours.
425: They should thus be representative of a wide class of models.
426:
427: I am thus proposing a new instability mechanism, by which the
428: halo will stimulate, rather than restrain, bar growth. This of course
429: will only work
430: if the halo is non-rigid and is capable of absorbing positive angular
431: momentum. Similarly the bar should also be non-rigid. Further analysis
432: of this mechanism will be given elsewhere.
433:
434: \acknowledgments
435:
436: I would like to thank A. Bosma, M. Tagger and F. Masset for
437: stimulating discussions, and A. Misiriotis and J.~C. Lambert for
438: their help with the software calculating the orbital frequencies. I
439: would also like to thank the Region PACA, the
440: INSU/CNRS, the University of Aix-Marseille I and the IGRAP for funds to develop
441: the GRAPE computing facilities used for the simulations discussed in
442: this paper.
443:
444: \begin{thebibliography}{}
445: \bibitem[Athanassoula (1984)]{atha84}
446: Athanassoula, E. 1984, Physics Rep., 114, 319
447: \bibitem[Athanassoula (1996)]{atha96}
448: Athanassoula, E. 1996, in IAU Colloq. 157, Barred Galaxies,
449: ed. R. Buta, D. A. Crocker \& B. G. Elmegreen (San Francisco ASP), 309
450: \bibitem[Athanassoula (2002a)]{atha02a}
451: Athanassoula, E. 2002a, in IAU Symp. 208, Astrophysical
452: Supercomputing Using Particle Simulations, eds. J. Makino \& P. Hut,
453: PASP, in press, and astro-ph/0112076
454: \bibitem[Athanassoula 2002b] {atha02b}
455: Athanassoula, E. 2002b, in ASP Conf. Ser., The dynamics, structure
456: and history of galaxies, eds. G. S. Da Costa \& E. M. Sadler,
457: in press and astro-ph/0112077
458: \bibitem [Athanassoula \& Misiriotis 2002]{am02}
459: Athanassoula, E., \& Misiriotis, A. 2002, \mnras, 330, 35 (AM)
460: \bibitem [Athanassoula \& Sellwood 1986]{as86}
461: Athanassoula, E., \& Sellwood, J. A. 1986, \mnras, 221, 213
462: \bibitem [Binney \& Spergel 1982]{bs}
463: Binney, J., \& Spergel, D. 1982, \apj, 252, 308
464: \bibitem [Combes et al 1990]{cdfp}
465: Combes, F., Debbasch, F., Friedli, D., \& Pfenniger, D. 1990, \aap,
466: 233, 82
467: \bibitem [Contopoulos \& Grosb{\o}l 1989] {cg89}
468: Contopoulos G., Grosb{\o}l P., 1989, \aapr, 1,261
469: \bibitem [Debattista \& Sellwood 1998]{ds98}
470: Debattista V. P., \& Sellwood, J. A. 1998, \apj, 493, L5
471: \bibitem [Debattista \& Sellwood 2000]{ds00}
472: Debattista V. P., \& Sellwood, J. A. 2000, \apj, 543, 704
473: \bibitem [Hernquist]{hern93}
474: Hernquist L., 1993, \apjs, 86, 389
475: \bibitem [Hernquist \& Weinberg]{hw92}
476: Hernquist L., \& Weinberg. M. D. 1992, \apj, 400, 80
477: \bibitem [Laskar 1990]{lask}
478: Laskar, J. 1990, Icarus, 88, 266
479: \bibitem [Little \& Carlberg]{lc}
480: Little, B. \& Carlberg, R. G. 1991, \mnras, 250, 161
481: \bibitem [Lynden-Bell \& Kalnajs 1972]{lk72}
482: Lynden-Bell, D., \& Kalnajs, A. J. 1972, \mnras, 157, 1
483: \bibitem [Ostriker \& Peebles 1973]{op73}
484: Ostriker, J. P., \& Peebles, P. J. E. 1973, \apj, 186, 467
485: \bibitem [Sellwood \& Wilkinson 1993]{SW}
486: Sellwood, J. A. \& Wilkinson, A. 1993, Rep. Prog. Phys., 56, 173
487: \bibitem [Tremaine \& Weinberg 1984]{tw84}
488: Tremaine, S., \& Weinberg, M. D. 1984, \mnras, 209, 729
489: \bibitem [Weinberg 1985]{w85}
490: Weinberg, M. D. 1985, \mnras, 213, 451
491: \end{thebibliography}
492:
493: \begin{figure}
494: \epsscale{0.85}
495: \plotone{f1.eps}
496: \caption{Basic results for the three simulations discussed
497: in the text. The left
498: panels correspond to model MD, the middle ones to model MH and the
499: right ones to model RH. The first and second row of panels give
500: the circular velocity curves at $t$ = 0 and 800, respectively
501: (solid lines). The contribution of the disc component is given by
502: a dashed line, and that of the halo by a dotted line.
503: The third row gives, again for $t$ = 800,
504: the isodensities of the disc component when seen
505: face-on. The size of the square box
506: for the third row of panels is 10 initial disc scale
507: lengths. The last row gives the relative amplitude of the $m$ = 2
508: (solid line), $m$ = 4 (dashed line), 6 (dot-dashed line) and 8 (dotted
509: line) components of the mass, again at $t$ = 800.
510: }
511: \label{fig:basic}
512: \end{figure}
513:
514: \clearpage
515:
516: \begin{figure}
517: \plotone{f2.eps}
518: \figcaption{Number density, $N_R$, of particles as a
519: function of the frequency ratio $R = (\Omega -
520: \Omega_p) / \kappa$, for simulation MD (upper panels) and MH (lower
521: panels) at $t$ = 800. The left panels correspond to the disc component
522: and the right ones to the halo. The dot-dashed vertical lines give
523: the positions of the main resonances. The results for the halo
524: component have been rescaled, so as to take into account the different
525: number of particles in the disc and halo components, and thus allow
526: immediate comparisons.
527: }
528: \label{fig:res}
529: \end{figure}
530:
531:
532:
533: %% If you are not including electonic art with your submission, you may
534: %% mark up your captions using the \figcaption command. See the
535: %% User Guide for details.
536: %%
537: \end{document}
538:
539: %%
540: %% End of file `sample.tex'.
541: