astro-ph0610535/ms.tex
1: %\documentclass[manuscript, flushrt]{aastex}
2: 
3: % The following generates a more compact, one-column version
4: \documentclass[12pt,preprint]{aastex}
5: %\documentclass[preprint]{aastex}
6: 
7: % The following is the two-column ApJ published form
8: %\documentclass{emulateapj}
9: %\usepackage{apjfonts}
10: 
11: %\usepackage[]{natbib}
12: 
13: % Define my comment style and Ignore
14: \long\def\Ignore#1{\relax}
15: \long\def\Comment#1{{\it\footnotesize #1}}
16: 
17: 
18: \newcommand{\Md}        {M_{\rm d}}
19: \newcommand{\Rd}        {R_{\rm d}}
20: \newcommand{\Mb}        {M_{\rm b}}
21: \newcommand{\Rb}        {R_{\rm b}}
22: \newcommand\degrees{^\circ}
23: 
24: \newcommand{\Msun}{\mbox{$\rm M_{\odot}$}}
25: \newcommand{\amp}[1]{\mbox{$A_{\rm #1}$}}
26: \newcommand{\tme}[1]{\mbox{$\tau_{\rm #1}$}}
27: \newcommand{\omg}[1]{\mbox{$\Omega_{\rm #1}$}}
28: \newcommand{\aep}[1]{\mbox{$a_{\epsilon,{\rm #1}}$}}
29: \newcommand{\lag}[1]{\mbox{$R_{\rm c,{#1}}$}}
30: 
31: \newcommand{\eg}{e.g.}
32: \newcommand{\etc}{etc.}
33: \newcommand{\ie}{i.e.}
34: \newcommand{\cf}{cf.}
35: 
36: 
37: \shorttitle{Double-Barred Galaxies From Pseudo-Bulges}
38:  
39: \shortauthors{Debattista \& Shen}
40: 
41: \begin{document}
42: \title{Long-Lived Double-Barred Galaxies From Pseudo-Bulges}
43: 
44: \author{Victor P. Debattista\altaffilmark{1}}
45:  \affil{Astronomy Department, University of Washington, Box 351580,
46: Seattle, WA 98195}
47: 
48: \altaffiltext{1}{Brooks Prize Fellow; email: debattis@astro.washington.edu}
49: 
50: \and
51: 
52: \author{Juntai Shen\altaffilmark{2}} 
53: \affil{McDonald Observatory, The University of
54: Texas at Austin, 1 University Station, C1402, Austin, TX 78712}
55: \altaffiltext{2}{Harlan J. Smith Fellow; email: shen@astro.as.utexas.edu}
56: 
57: %\slugcomment{Draft Version of  \textit{\today}}
58: 
59: \begin{abstract}
60: A large fraction of barred galaxies host secondary bars that are
61: embedded in their large-scale primary counterparts. These are common
62: also in gas poor early-type barred galaxies.  The evolution of such
63: double-barred galaxies is still not well understood, partly because of
64: a lack of realistic $N$-body models with which to study them.  Here we
65: report a new mechanism for generating such systems, namely the
66: presence of rotating pseudo-bulges.  We demonstate with high mass and
67: force resolution collisionless $N$-body simulations that long-lived
68: secondary bars can form spontaneously without requiring gas, contrary
69: to previous claims.  We find that secondary bars rotate faster than
70: primary ones. The rotation is not, however, rigid: the secondary bars
71: pulsate, with their amplitude and pattern speed oscillating as they
72: rotate through the primary bars.  This self-consistent study supports
73: previous work based on orbital analysis in the potential of two
74: rigidly rotating bars.  The pulsating nature of secondary bars may
75: have important implications for understanding the central region of
76: double-barred galaxies.
77: \end{abstract}
78: 
79: \keywords{stellar dynamics --- galaxies: bulges --- galaxies:
80: evolution --- galaxies: formation --- galaxies: kinematics and
81: dynamics --- galaxies: structure}
82: 
83: 
84: \section{Introduction} 
85: \label{sec:intro}
86: 
87: Double-barred (S2B) galaxies, consisting of a small scale
88: nuclear/secondary bar (B2) embedded within a large scale primary bar
89: (B1), have been known for over thirty years \citep[e.g.][]{devauc_75}.
90: \citet{erw_spa_02} carefully compiled statistics for early-type
91: optically-barred galaxies from images by both the WIYN telescope and
92: the {\it Hubble Space Telescope}, and concluded that at least one
93: quarter of them are double-barred.  The facts that inner bars are also
94: seen in near-infrared \citep[e.g.][]{mul_etal_97,lai_etal_02}, and
95: that gas-poor S0s often contain inner bars indicate that most of them
96: are stellar structures.  S2Bs may play an important role in the
97: formation and nurture of supermassive black holes (SMBHs).  S2B
98: galaxies have been hypothesized to be a possible mechanism for driving
99: gas past the inner Lindblad resonance of B1s, feeding SMBHs and
100: powering AGN \citep{shl_etal_89}.  S2Bs have also been suggested as a
101: mechanism for forming SMBHs directly \citep{beg_etal_06}.
102: 
103: Such fueling requires that the B2 and the B1 are dynamically
104: decoupled.\footnote{In this context, by decoupled we mean only that
105: $\omg{B2} \neq \omg{B1}$, where \omg{B2} (\omg{B1}) is the pattern
106: speed of the B2 (B1).}  The random apparent orientations of B1s and
107: B2s in nearly face-on galaxies points to dynamical decoupling
108: \citep{but_cro_93, fri_mar_93}.  But images alone cannot reveal how
109: the two bars rotate through each other.  Kinematic evidence of
110: decoupling, using either gas or stars, is harder to obtain
111: \citep{pet_wil_02, sch_etal_2002,moi_etal_04}.  Indirect evidence for
112: decoupling was suggested by \citet{ems_etal_01} based on rotation
113: velocity peaks inside the B2s in three S2B galaxies.  Conclusive
114: direct kinematic evidence for a decoupled B2 was obtained for NGC 2950
115: by \citet{cor_deb_agu_03} who showed, using the method of
116: \citet{tre_wei_84}, that its B1 and B2 cannot be rotating at the same
117: rate.
118: 
119: An important advance in understanding S2B galaxies came from the
120: development by \citet[][hereafter MS00]{mac_spa_97,mac_spa_00} of the
121: formalism necessary for studying their orbits.  They introduced the
122: concept of loops, families of orbits in which particles return to the
123: same curve, but not the same position, after the two bars return to
124: the same relative orientation.  MS00 considered two models assuming
125: that the B2 is more rapidly rotating than the B1: the B2 in
126: their Model 1 ended near its corotation radius, while in their Model 2
127: it ended well inside this radius.  MS00 were unable to find loop
128: orbits supporting the outer parts of the B2 in Model 1 but
129: succeeded in doing so in the more slowly rotating Model 2.  Using
130: hydrodynamical simulations of such slowly rotating rigid B2s
131: \citet{mac_etal_02} found them to be inefficient at driving gas to
132: small radii.
133: 
134: Such models are not fully self-consistent since, in general, nested
135: bars cannot rotate rigidly through each other \citep{lou_ger_88}.  In
136: fact non-solid body rotation was hinted at by the loop orbit
137: calculations of MS00.
138: %
139: $N$-body simulations provide one route to more self-consistent models
140: of S2Bs, but until now there existed a paucity of such models.  Most
141: numerical studies \citep[\eg][]{shl_hel_02, fri_mar_93, eng_shl_04}
142: required gas to form B2s; for example, \citet{hel_etal_01} formed them
143: via viscosity-driven instabilities in nuclear gas rings, which lead to
144: B2s rotating slower than B1s.  But the presence of B2s in a large
145: fraction of gas-poor early-type galaxies \citep{erw_spa_02,pet_wil_02}
146: indicates that B2s are not an exclusively gas dynamical phenomenon.
147: Counter-rotation in stellar disks can lead to decoupled
148: counter-rotating bars \citep{sel_mer_94, friedl_96, dav_hun_97}, but
149: such counter-rotation is infrequent \citep{kui_etal_96}.  Only
150: Rautiainen and collaborators \citep{rau_sal_99, rau_etal_02} have
151: succeeded in forming long-lived B2s rotating in the same sense as the
152: B1 in purely collisionless studies.  The mechanism by which the B2s
153: formed in these however simulations remains unclear.
154: 
155: In light of the increasing evidence that SMBH feedback may play an
156: important role in galaxy formation \citep{spr_etal_05b} and the
157: possibility that S2Bs may provoke AGNs, the paucity of self-consistent
158: $N$-body models of S2Bs is a major hindrance to further theoretical
159: development.
160: %
161: The time is ripe, therefore, to examine whether unambiguous and
162: independently rotating nested bars can form in high resolution
163: collisionless simulations.  \citet{kor_ken_04} pointed out that a
164: nuclear bar constitutes strong evidence of a pseudo-bulge, \ie, a
165: bulge with a disky character.  Such pseudo-bulges form through the
166: secular evolution of disks, via both gas and stellar dynamical
167: processes \citep[see the review of ][]{kor_ken_04}.  One of the main
168: characteristics of pseudo-bulges is that they rotate rapidly, a
169: property which favors them to become bar unstable.  In this work, we
170: demonstrate that a rapidly rotating bulge can develop a long-lived B2
171: in collisionless $N$-body simulations.
172: 
173: 
174: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
175: 
176: 
177: \section{Model Setup}
178: \label{sec:setup}
179: 
180: We focus on two examples of simulations which formed long-lasting
181: double-barred systems taken from a large survey of such
182: simulations. Our high-resolution simulations consist of live disk and
183: bulge components in a rigid halo potential.  We restrict ourselves to
184: rigid halos to afford high mass resolution in the nuclear regions, to
185: study the complicated co-evolution of the two bars without the
186: additional evolution introduced by the halo, and to compare with the
187: models of MS00.  We defer the study of S2B systems in live halos to a
188: future publication.  The rigid halos used in this study are all
189: logarithmic potentials $\Phi(r) = \frac{1}{2}V_{\rm h}^2~ \ln(r^2 +
190: r_{\rm h}^2)$.  We set $V_{\rm h} = 0.6$ in both runs and $r_{\rm h} =
191: 15$ in run 1 and $r_{\rm h} = 10$ in run 2.
192: %
193: Both initial disks in our simulations have exponential surface
194: densities with scale-length $\Rd$, mass $\Md$ and Toomre-$Q\simeq
195: 2$. The bulge was generated using the method of \citet{pre_tom_70} as
196: described in \citet{deb_sel_00}, where a distribution function is
197: integrated iteratively in the global potential, until convergence.  In
198: both cases the bulge has mass $\Mb=0.2\Md$ and we used an isotropic
199: King model distribution function.  The bulge truncation radius is
200: $0.7\Rd$ in run 1 and $1.0\Rd$ in run 2.  The bulge set up this way is
201: non-rotating.  We introduce bulge rotation by simply reversing the
202: velocities of bulge particles with negative angular momenta, which is
203: still a valid solution of the collisionless Boltzmann equation
204: \citep{lynden_62}.  The bulge in run 1 is flattened by the disk
205: potential to an edge-on projected ellipticity of $\epsilon_{\rm b}
206: \simeq 0.25$.  The ratio $V_p/{\bar\sigma} \simeq 0.8$, where $V_p$ is
207: the peak velocity and ${\bar\sigma}$ is the average velocity
208: dispersion inside the half mass radius.  In run 2, the corresponding
209: values are $\epsilon_{\rm b} \simeq 0.38$ and $V_p/{\bar\sigma} \simeq
210: 0.7$.  The kinematic values relative to the oblate isotropic rotators
211: are $(V_p/{\bar\sigma})_* \simeq 1.3$ for run 1 and
212: $(V_p/{\bar\sigma})_* \simeq 0.9$ for run 2.  Thus both pseudo-bulges
213: are above or close to the locus of oblate isotropic rotators.  These
214: pseudo-bulges are less tangentially biased and more pressure supported
215: than rotationally-supported pseudo-bulges which would form out of gas
216: driven to small radii.  Our simulations therefore probably
217: under-estimate the tendency for pseudo-bulges to form nuclear bars.
218: 
219: We use $\Rd$ and $\Md$ as the units of length and mass, respectively,
220: and the time unit is $(\Rd^3/G\Md)^{1/2}$.  If we scale these units to
221: the physical values $\Md = 2.3 \times 10^{10} \Msun$ and $\Rd = 2.5$
222: kpc, then a unit of time is $12.3$ Myr.  We use a force resolution
223: (softening) of $0.01$, which scaled to the above physical units
224: corresponds to 25 pc.  Both models had $1.2 \times 10^6$ equal mass
225: particles, with $10^6$ in the disk.  These simulations were evolved
226: with a 3-D cylindrical polar grid code \citep{sel_val_97}.  This code
227: expands the potential in a Fourier series in the cylindrical polar
228: angle $\phi$; we truncated the expansion at $m=8$. Forces in the
229: radial direction are solved for by direct convolution with the Greens
230: function while the vertical forces are obtained by fast Fourier
231: transform.
232: %
233: We used grids measuring $N_R\times N_\phi \times N_z = 58 \times 64
234: \times 375$.  The vertical spacing of the grid planes was $\delta z =
235: 0.01 R_{\rm d}$.  Time integration used a leapfrog integrator with a
236: fixed time-step $\delta t = 0.04$.
237: 
238: 
239: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
240: 
241: 
242: \section{Results}
243: \label{sec:results}
244: 
245: Figure~\ref{fig:snapshots} gives a general view of the evolution of
246: run 1 over 750 time units ($\sim 9.2$ Gyr in our standard scaling).  A
247: nuclear bar forms rapidly (before $t=10$), as the dynamical times in
248: the inner galaxy are much shorter than in the outer part. The pattern
249: speed, \omg{B2}, of this nuclear bar is large at this stage, and it
250: extends to nearly its corotation radius, \lag{B2}, indicating that it
251: forms by the usual bar instability \citep{toomre_81}.  The B1 forms at
252: a later stage, between $t=100$ and $200$.  The evolution of the
253: amplitudes of the B1 and B2 (\amp{B1} and \amp{B2} respectively),
254: defined as the Fourier $m=2$ amplitude over the radial ranges $ 0.5
255: \leq R \leq 2$ and $R \leq 0.3$, is shown in the left panel of Figure
256: \ref{fig:barampl}.  The B2 is strong initially, but it weakens once
257: the B1 forms.  At the same time $\omg{B2}$ also decreases and, since
258: its semi-major axis does not change substantially, it no longer
259: extends to \lag{B2}.
260: %
261: The transition to a stable S2B state is accomplished during a
262: seemingly chaotic period during which both bars undergo phases when
263: they are rather weak.  After $t=250$, however, the B2 settles into an
264: oscillatory steady state with \amp{B2} exhibiting regular oscillations.
265: The double-barred state persists to the end of the simulation, lasting
266: for $\sim 7$ Gyr. The B2 shows up in both the disk and bulge particles.
267: 
268: The B2 is stronger when the bars are perpendicular, and weaker when
269: they are parallel to each other (Figure \ref{fig:barampl}). Note that
270: this behavior is exactly opposite to the variations of gaseous rings
271: in \citet{hel_etal_01}. The amplitude of the primary bar instead varies
272: in the opposite sense with respect to the relative phase of the two
273: bars, although the amplitude of this oscillation is smaller.  \amp{B1}
274: also decreases slowly after $t=250$, possibly because orbits
275: supporting the B1 are gradually disrupted by the relatively strong
276: inner bar.  As a consequence, the oscillations in \amp{B2} decrease as
277: the B1 weakens.
278: 
279: 
280: The right panel of Figure \ref{fig:barampl} shows the evolution of
281: \amp{B1} and \amp{B2} in run 2.  The main difference between run 2 and
282: run 1 is that the initial bulge is larger in run 2, allowing the B2 to
283: dominate the global dynamics.  As a result, the B1 oscillates more
284: strongly than the B2.  This probably represents an extreme case of the
285: dynamical influence of a B2 on a B1. 
286: 
287: The long-lived B2 rotates faster than the B1: between $t=300$ and
288: $t=400$ the average rotation period of the B2 in run 1 is about
289: $\tme{B2}\simeq 17.6$, and for the B1 $\tme{B1} \simeq 27.8$.  The
290: pattern speed of the B2, \omg{B2}, also varies with the relative phase
291: of the two bars: it is larger when the two bars align, and smaller
292: when they are orthogonal. We plot in Figure \ref{fig:crframe} the
293: system in the corotating frame of the B1.  The variations of both
294: \amp{B2} and \omg{B2} are readily visible. The variation of \omg{B2}\ can
295: be $>20\%$ but is much less significant for \omg{B1} (Figure
296: \ref{fig:barphase}).  Defining $\left< \omg{B2}\right>$ ($\left<
297: \omg{B1}\right>$) as the average pattern speed of the B2 (B1) over one
298: relative rotation, we plot in the inset of Figure \ref{fig:barphase}
299: the phase difference between $\left< \omg{B2}\right>t$ ($\left<
300: \omg{B1}\right>t$) and the phase of the B2 (B1).  The B1 is seen to
301: rotate with a rather constant \omg{B1}\ but the B2 experiences a large
302: variation in \omg{B2} over one relative rotation.
303: 
304: 
305: Figure \ref{fig:ellipsefit} presents ellipse fits using {\sc iraf} for
306: times when the B2 and B1 are perpendicular and at $\sim 45\degrees$ to
307: each other.  In both cases the phase of the B2 is constant to within
308: $10\degrees$ and there is little sense of spirality in it.  This is
309: distinctly a nuclear bar rather than a nuclear spiral.
310: %
311: 
312: We measured the sizes of the two bars, for two different relative
313: orientations at $t=340$ and at $t=350$, as the larger radius where the
314: bar phase deviates by more than $10\degrees$ from a constant value.
315: We find a semi-major axis ratio $\simeq 0.10$ ($\simeq 0.12$) at
316: $t=340$ ($t=350$), in good agreement with the typical size ratio of local
317: S2B systems \citep{erw_spa_02}.
318: 
319: 
320: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
321: 
322: 
323: \section{Discussion and Conclusions}
324: \label{sec:discussion}
325: 
326: Our self-consistent simulations of S2B systems can be compared to the
327: models of MS00.  The simulations all exhibited oscillating pattern
328: speeds and amplitudes for one or both bars.  Similarly MS00 found that
329: the $x_2$ loops supporting the B2 change axis ratios and lead or trail
330: the rigid figure of the B2, as the bars rotate.  The loop orbits of
331: MS00 were more elongated in the B2 region when the two bars were
332: orthogonal than when they were parallel.  The pulsating character of
333: the self-consistent B2 in the simulations provides strong evidence
334: that $x_2$ loops are the backbone of the double bars in these
335: simulations.  This behavior is also in good agreement with the earlier
336: prediction by \citet{lou_ger_88} that independent rigid rotation of
337: two bars is not possible.
338: %
339: The $x_2$ loop orbits of MS00 also suggested that \omg{B2} would be
340: largest when the two bars are parallel, which is also borne out by the
341: simulations.  Furthermore, MS00 were unable to find supporting $x_2$
342: orbits when the B2 extended to about \lag{B2}, while we found that
343: \omg{B2} had to decrease once the B1 formed and the B2 did not extend
344: to \lag{B2}, again in good agreement with MS00. 
345: %
346: Our simulations also suggest that observationally there should be a
347: slight excess of close-to-perpendicular double bars, as the secondary
348: bar tends to rotate slower when two bars are perpendicular.
349: 
350: The main objective of this work is to create S2B systems and show how
351: the two bars form spontaneously, interact and evolve.  Our simulations
352: all form B2s before they form B1s.  However this is not a prediction
353: of our model and it occurs only because, for simplicity, we introduced
354: our rotating pseudo-bulge from $t=0$.  It is more likely that a
355: pseudo-bulge would form after gas is driven to the center by a
356: pre-existing B1.  Our pseudo-bulges all had rotation; B2s did not form
357: in simulations without pseudo-bulge rotation \citep[these results will
358: be presented elsewhere, but see for example][]{debatt_03}.  In
359: contrast, \citet{rau_sal_99} produced B2s even though their bulges
360: were analytic.
361: 
362: We are able to form well-resolved, long-lived B2s in purely
363: collisionless $N$-body simulations.  The nuclear bars are distinctly
364: barred, not spiral, and reach to the center.  These simulations
365: demonstrate that B2s do not need to be gaseous.  We confirm that
366: pseudo-bulge rotation may be an important ingredient for the formation
367: of double-barred galaxies \citep{kor_ken_04}. The required degree of
368: rotation is modest and not greater than that associated with
369: pseudo-bulges \citep{kormen_93}.  The B2s in these simulations rotate
370: faster than the B1s.  The implications of the pulsating nature of B1s
371: on central gas inflow are unclear at present.
372: %
373: This new method for forming S2B models reliably and repeatedly should
374: prove a boon to exploring their dynamics and evolution, their
375: observational properties, their effect on gas, \etc\ We will report on
376: these issues elsewhere.
377: 
378: 
379: %     A C K N O W L E D G E M E N T
380: \acknowledgements 
381: V.P.D. is supported by a Brooks Prize Fellowship at the
382: University of Washington and receives partial support from NSF ITR
383: grant PHY-0205413.  V.P.D. thanks the University of Texas at Austin
384: for hospitality during part of this project.
385: 
386: %     R E F E R E N C E S
387: \begin{thebibliography}{}
388: 
389: \bibitem[\protect\citeauthoryear{{Begelman}, {Volonteri}, \& {Rees}}{{Begelman}
390:   et~al.}{2006}]{beg_etal_06}
391: {Begelman}, M.~C., {Volonteri}, M.,  \& {Rees}, M.~J. 2006, astro-ph/0602363
392: 
393: \bibitem[\protect\citeauthoryear{{Buta} \& {Crocker}}{{Buta} \&
394:   {Crocker}}{1993}]{but_cro_93}
395: {Buta}, R.,  \& {Crocker}, D.~A. 1993, \aj, 105, 1344
396: 
397: \bibitem[\protect\citeauthoryear{{Corsini}, {Debattista}, \&
398:   {Aguerri}}{{Corsini} et~al.}{2003}]{cor_deb_agu_03}
399: {Corsini}, E.~M., {Debattista}, V.~P.,  \& {Aguerri}, J.~A.~L. 2003, \apjl,
400:   599, L29
401: 
402: \bibitem[\protect\citeauthoryear{{Davies} \& {Hunter}}{{Davies} \&
403:   {Hunter}}{1997}]{dav_hun_97}
404: {Davies}, C.~L.,  \& {Hunter}, J.~H. 1997, \apj, 484, 79
405: 
406: \bibitem[\protect\citeauthoryear{{de Vaucouleurs}}{{de
407:   Vaucouleurs}}{1975}]{devauc_75}
408: {de Vaucouleurs}, G. 1975, \apjs, 29, 193
409: 
410: \bibitem[\protect\citeauthoryear{{Debattista}}{{Debattista}} {2003}]{debatt_03}
411: {Debattista}, V.~P. 2003, \mnras, 342, 1194
412: 
413: \bibitem[\protect\citeauthoryear{{Debattista} \& {Sellwood}}{{Debattista} \&
414:   {Sellwood}}{2000}]{deb_sel_00}
415: {Debattista}, V.~P.,  \& {Sellwood}, J.~A. 2000, \apj, 543, 704
416: 
417: \bibitem[\protect\citeauthoryear{{Emsellem} et~al.}{{Emsellem}
418:   et~al.}{2001}]{ems_etal_01}
419: {Emsellem}, E., {Greusard}, D., {Combes}, F., {Friedli}, D., {Leon}, S.,
420:   {P{\'e}contal}, E.,  \& {Wozniak}, H. 2001, \aap, 368, 52
421: 
422: \bibitem[\protect\citeauthoryear{{Englmaier} \& {Shlosman}}{{Englmaier} \&
423:   {Shlosman}}{2004}]{eng_shl_04}
424: {Englmaier}, P.,  \& {Shlosman}, I. 2004, \apjl, 617, L115
425: 
426: \bibitem[\protect\citeauthoryear{{Erwin}}{{Erwin}}{2004}]{erwin_04}
427: {Erwin}, P. 2004, \aap, 415, 941
428: 
429: \bibitem[\protect\citeauthoryear{{Erwin} \& {Sparke}}{{Erwin} \&
430:   {Sparke}}{2002}]{erw_spa_02}
431: {Erwin}, P.,  \& {Sparke}, L.~S. 2002, \aj, 124, 65
432: 
433: \bibitem[\protect\citeauthoryear{{Friedli}}{{Friedli}}{1996}]{friedl_96}
434: {Friedli}, D. 1996, \aap, 312, 761
435: 
436: \bibitem[\protect\citeauthoryear{{Friedli} \& {Martinet}}{{Friedli} \&
437:   {Martinet}}{1993}]{fri_mar_93}
438: {Friedli}, D.,  \& {Martinet}, L. 1993, \aap, 277, 27
439: 
440: \bibitem[\protect\citeauthoryear{{Heller}, {Shlosman}, \& {Englmaier}}{{Heller}
441:   et~al.}{2001}]{hel_etal_01}
442: {Heller}, C., {Shlosman}, I.,  \& {Englmaier}, P. 2001, \apj, 553, 661
443: 
444: \bibitem[\protect\citeauthoryear{{Kormendy}}{{Kormendy}}{1993}]{kormen_93}
445: {Kormendy}, J. 1993, in IAU Symp. 153: Galactic Bulges, ed. H.~{Dejonghe} \&
446:   H.~J. {Habing}, 209
447: 
448: \bibitem[\protect\citeauthoryear{{Kormendy} \& {Kennicutt}}{{Kormendy} \&
449:   {Kennicutt}}{2004}]{kor_ken_04}
450: {Kormendy}, J.,  \& {Kennicutt}, R.~C. 2004, \araa, 42, 603
451: 
452: \bibitem[\protect\citeauthoryear{{Kuijken}, {Fisher}, \&
453:   {Merrifield}}{{Kuijken} et~al.}{1996}]{kui_etal_96}
454: {Kuijken}, K., {Fisher}, D.,  \& {Merrifield}, M.~R. 1996, \mnras, 283, 543
455: 
456: \bibitem[\protect\citeauthoryear{{Laine} et~al.}{{Laine}
457:   et~al.}{2002}]{lai_etal_02}
458: {Laine}, S., {Shlosman}, I., {Knapen}, J.~H.,  \& {Peletier}, R.~F. 2002, \apj,
459:   567, 97
460: 
461: \bibitem[\protect\citeauthoryear{{Louis} \& {Gerhard}}{{Louis} \&
462:   {Gerhard}}{1988}]{lou_ger_88}
463: {Louis}, P.~D.,  \& {Gerhard}, O.~E. 1988, \mnras, 233, 337
464: 
465: \bibitem[\protect\citeauthoryear{{Lynden-Bell}}{{Lynden-Bell}}{1962}]{lynden_62}
466: {Lynden-Bell}, D. 1962, \mnras, 123, 447
467: 
468: \bibitem[\protect\citeauthoryear{{Maciejewski} \& {Sparke}}{{Maciejewski} \&
469:   {Sparke}}{1997}]{mac_spa_97}
470: {Maciejewski}, W.,  \& {Sparke}, L.~S. 1997, \apjl, 484, L117
471: 
472: \bibitem[\protect\citeauthoryear{{Maciejewski} \& {Sparke}}{{Maciejewski} \&
473:   {Sparke}}{2000}]{mac_spa_00}
474: {Maciejewski}, W.,  \& {Sparke}, L.~S. 2000, \mnras, 313, 745
475: 
476: \bibitem[\protect\citeauthoryear{{Maciejewski} et~al.}{{Maciejewski}
477:   et~al.}{2002}]{mac_etal_02}
478: {Maciejewski}, W., {Teuben}, P.~J., {Sparke}, L.~S.,  \& {Stone}, J.~M. 2002,
479:   \mnras, 329, 502
480: 
481: \bibitem[\protect\citeauthoryear{{Moiseev}, {Vald{\'e}s}, \&
482:   {Chavushyan}}{{Moiseev} et~al.}{2004}]{moi_etal_04}
483: {Moiseev}, A.~V., {Vald{\'e}s}, J.~R.,  \& {Chavushyan}, V.~H. 2004, \aap, 421,
484:   433
485: 
486: \bibitem[\protect\citeauthoryear{{Mulchaey}, {Regan}, \& {Kundu}}{{Mulchaey}
487:   et~al.}{1997}]{mul_etal_97}
488: {Mulchaey}, J.~S., {Regan}, M.~W.,  \& {Kundu}, A. 1997, \apjs, 110, 299
489: 
490: \bibitem[\protect\citeauthoryear{{Petitpas} \& {Wilson}}{{Petitpas} \&
491:   {Wilson}}{2002}]{pet_wil_02}
492: {Petitpas}, G.~R.,  \& {Wilson}, C.~D. 2002, \apj, 575, 814
493: 
494: \bibitem[\protect\citeauthoryear{{Prendergast} \& {Tomer}}{{Prendergast} \&
495:   {Tomer}}{1970}]{pre_tom_70}
496: {Prendergast}, K.~H.,  \& {Tomer}, E. 1970, \aj, 75, 674
497: 
498: \bibitem[\protect\citeauthoryear{{Rautiainen}, \& {Salo}}{{Rautiainen}
499: \& {Salo}}{1999}]{rau_sal_99} {Rautiainen}, P., \& {Salo}, H.  1999,
500: \aap, 348, 737
501: 
502: \bibitem[\protect\citeauthoryear{{Rautiainen}, {Salo}, \&
503:   {Laurikainen}}{{Rautiainen} et~al.}{2002}]{rau_etal_02}
504: {Rautiainen}, P., {Salo}, H.,  \& {Laurikainen}, E. 2002, \mnras, 337, 1233
505: 
506: \bibitem[\protect\citeauthoryear{{Schinnerer} et~al.}{{Schinnerer}
507:   et~al.}{2002}]{sch_etal_2002}
508: {Schinnerer}, E., {Maciejewski}, W., {Scoville}, N.,  \& {Moustakas}, L.~A.
509:   2002, \apj, 575, 826
510: 
511: \bibitem[\protect\citeauthoryear{{Sellwood} \& {Merritt}}{{Sellwood} \&
512:   {Merritt}}{1994}]{sel_mer_94}
513: {Sellwood}, J.~A.,  \& {Merritt}, D. 1994, \apj, 425, 530
514: 
515: \bibitem[\protect\citeauthoryear{{Sellwood} \& {Valluri}}{{Sellwood} \&
516:   {Valluri}}{1997}]{sel_val_97}
517: {Sellwood}, J.~A.,  \& {Valluri}, M. 1997, \mnras, 287, 124
518: 
519: \bibitem[\protect\citeauthoryear{{Shlosman}, {Frank}, \& {Begelman}}{{Shlosman}
520:   et~al.}{1989}]{shl_etal_89}
521: {Shlosman}, I., {Frank}, J.,  \& {Begelman}, M.~C. 1989, \nat, 338, 45
522: 
523: \bibitem[\protect\citeauthoryear{{Shlosman} \& {Heller}}{{Shlosman} \&
524:   {Heller}}{2002}]{shl_hel_02}
525: {Shlosman}, I.,  \& {Heller}, C.~H. 2002, \apj, 565, 921
526: 
527: \bibitem[\protect\citeauthoryear{{Silverman}}{{Silverman}}{1986}]{silver_86}
528: {Silverman}, B.~W. 1986, {Density estimation for statistics and data analysis}
529:   (Monographs on Statistics and Applied Probability, London: Chapman and Hall)
530: 
531: \bibitem[\protect\citeauthoryear{{Springel}, {Di Matteo}, \&
532:   {Hernquist}}{{Springel} et~al.}{2005}]{spr_etal_05b}
533: {Springel}, V., {Di Matteo}, T.,  \& {Hernquist}, L. 2005, \mnras, 361, 776
534: 
535: \bibitem[\protect\citeauthoryear{{Toomre}}{{Toomre}}{1981}]{toomre_81}
536: {Toomre}, A. 1981, in Structure and Evolution of Normal Galaxies, ed. S.~M.,
537:   Fall \& D. Lynden-Bell (Cambridge: Cambridge University Press), 111
538: 
539: \bibitem[\protect\citeauthoryear{{Tremaine} \& {Weinberg}}{{Tremaine} \&
540:   {Weinberg}}{1984}]{tre_wei_84}
541: {Tremaine}, S.,  \& {Weinberg}, M.~D. 1984, \apjl, 282, L5
542: 
543: \end{thebibliography}
544: 
545: \clearpage
546: 
547: \begin{figure}
548: \centerline{
549: \includegraphics[angle=-90.,width=\hsize]{f1.ps}}
550: \caption{Images of total stellar distribution at various times, with
551: iso-density contours superposed. The contours are logarithmic and
552: separated by 0.2 dex.  The heavy short and long straight lines mark
553: the major axes of the B2 and B1, respectively.  The surface density is
554: obtained by smoothing every particle with an adaptive kernel
555: \citep{silver_86}.  Note that 100 time units is about 1.2 Gyr, and the
556: length unit is the scale-length of the initial disk.}
557: \label{fig:snapshots}
558: \end{figure}
559: 
560: \clearpage
561: 
562: \begin{figure}
563: \centerline{
564: \includegraphics[angle=0.,width=0.45\hsize]{f2a.ps}
565: \includegraphics[angle=0.,width=0.45\hsize]{f2b.ps}}
566: \caption{The time evolution of the bar amplitude of the B2 ({\it top
567:   panels}) and the B1 ({\it bottom panels}). In the insets, the dashed
568:   lines mark times when the two bars nearly align, while the dotted
569:   lines mark the time when they are perpendicular to each other. The
570:   beat period $\tme{\rm beat}=\tme{B1}\tme{B2}/2(\tme{B1}-\tme{B1})$. Run 1 is
571:   shown on the left, while run 2 is on the right.}
572: \label{fig:barampl}
573: \end{figure}
574: 
575: \clearpage
576: 
577: \begin{figure}
578: \centerline{
579: \includegraphics[angle=-90., width=\hsize]{f3.ps}}
580: \caption{The non-uniform relative rotation of the B2 for roughly half
581:   of a period, in the corotating frame of the B1 which remains
582:   horizontal.  The panels are equally-spaced in time. The straight
583:   line marks the major axis of the B2.  The B2 rotates faster when the
584:   two bars align than when the two bars are perpendicular.}
585: \label{fig:crframe}
586: \end{figure}
587: 
588: \clearpage
589: 
590: \begin{figure}
591: \centerline{
592: \includegraphics[angle=-90.,width=\hsize]{f4.ps}}
593: \caption{The time evolution of the phase of the B2, measured relative
594: to $t=300$. The dashed straight line is the least-square fit which
595: gives the slope $\left<\omg{B2}\right>$. The inset figure shows the
596: phase difference, $\Delta\phi$ between the phases of the bars and
597: $\left<\omg{}\right>t$, where $\left<\omg{}\right>$ is the pattern
598: speed averaged over one relative rotation of the two bars; the thick
599: line is for the B2 while the thin line is for the B1.}
600: \label{fig:barphase}
601: \end{figure}
602: 
603: \clearpage
604: 
605: \begin{figure}
606: \centerline{
607: \includegraphics[angle=-90.,width=0.5\hsize]{f5a.ps}
608: \includegraphics[angle=-90.,width=0.5\hsize]{f5b.ps}}
609: \centerline{
610: \includegraphics[angle=0.,width=0.5\hsize]{f5c.ps}
611: \includegraphics[angle=0.,width=0.5\hsize]{f5d.ps}}
612: \caption{Results of ellipse fits using {\sc iraf}.  Top panels:
613: snapshots of run 1 at $t=340$ (left) and $t=350$ (right).  The two
614: bars are at $\sim 45\degrees$ at $t=340$ and perpendicular at $t=350$.
615: Middle panels: Ellipticity as a function of semi-major axis (SMA) of
616: fitted ellipses.  Bottom panels: Position angle as a function of SMA
617: of fitted ellipses.}
618: \label{fig:ellipsefit}
619: \end{figure}
620: 
621: \end{document}
622: