1: %\documentclass[12pt]{aastex}
2: %\documentclass[12pt, preprint]{aastex}
3: \documentclass[natbib]{emulateapj}
4: %\usepackage{natbib}
5: %\usepackage{graphicx} % Add graphics capabilities
6: %\usepackage{epstopdf} % to include .eps graphics files with pdfLaTeX
7:
8: \newcommand{\myemail}{roskar@astro.washington.edu}
9:
10: \begin{document}
11:
12: \title{Riding the Spiral Waves: Implications of Stellar Migration for the Properties of Galactic Disks}
13:
14: \author{Rok Ro\v{s}kar\altaffilmark{1}, Victor
15: P. Debattista\altaffilmark{2}, Thomas R. Quinn\altaffilmark{1},
16: Gregory S. Stinson\altaffilmark{3}, James Wadsley\altaffilmark{3}}
17:
18: \altaffiltext{1}{Astronomy Department, University of Washington, Box
19: 351580, Seattle, WA 98195, USA {\tt
20: roskar;trq@astro.washington.edu}}
21: \altaffiltext{2}{RCUK Fellow at Centre for Astrophysics, University of Central
22: Lancashire, Preston, PR1 2HE, UK {\tt vpdebattista@uclan.ac.uk}}
23: \altaffiltext{3}{Department of Physics and Astronomy, McMaster University,
24: Hamilton, ON, L8S 4M1, Canada {\tt stinson;wadsley@mcmaster.ca}}
25:
26:
27: \begin{abstract}
28:
29: Stars in disks of spiral galaxies are usually assumed to remain roughly at their birth radii.
30: This assumption is built into decades of modelling of the
31: evolution of stellar populations in our own Galaxy and in external systems.
32: We present results from self-consistent high-resolution $N$-body + Smooth
33: Particle Hydrodynamics simulations of disk formation, in which stars migrate
34: across significant galactocentric distances due to resonant scattering with transient spiral arms,
35: while preserving their circular orbits. We investigate the implications
36: of such migrations for observed stellar populations. Radial migration provides an
37: explanation for the observed flatness and spread in the age-metallicity
38: relation and the relative lack of metal poor stars in the solar neighborhood.
39: The presence of radial migration also prompts rethinking of interpretations
40: of extra-galactic stellar population data, especially for determinations of star formation histories.
41:
42: \end{abstract}
43:
44: \keywords{galaxies: evolution --- galaxies: spiral --- galaxies: stellar content
45: --- Galaxy: solar neighborhood --- Galaxy: stellar content --- stellar dynamics}
46:
47: \section{Introduction}
48:
49: In a universe where mass assembles by accretion of progressively larger constituents,
50: thin disks of galaxies form during quiescent evolution following the major merging epoch
51: \citep{brook:2004,Robertson:2006}. The inner parts, with lower angular momentum
52: assemble first, followed by the higher angular momentum components, resulting in
53: ``inside-out'' build up of disk material \citep{Larson:1976,white:1991,Munoz-Mateos:2007}.
54: Because gas can efficiently dissipate energy, its motion around the galaxy is mostly circular.
55: Stars born in such gas disks are initially on nearly circular orbits, but their infant kinematic state is
56: highly fragile; non-axisymetric perturbations such as bars or spiral arms readily drive
57: the orbits away from circular. Despite the increase in eccentricity, mean orbital radii are
58: assumed to remain relatively constant. As a result, interpretations of Galactic and
59: extra-galactic stellar population observations invariably make the fundamental
60: assumption that stars remain at roughly their birth radii throughout their lives.
61: Consequently, this assumption is built into the vast majority of models of formation
62: and evolution of galactic disks over the past few decades \citep[e.g.][]{Tinsley:1975,
63: Matteucci:1989, Carigi:1996,Chiappini:1997,Boissier:1999}.
64:
65: Recent theoretical and observational evidence challenges this static picture.
66: Spirals have long been known to be an important source of
67: kinematic heating in galactic disks, gradually increasing the eccentricities of
68: stellar orbits \citep{Jenkins:1990}. However, if a star is
69: caught in the corotation resonance of a transient spiral it may move inward
70: or outward in radius while preserving the circularity of its orbit
71: \citep[][hereafter SB02]{Sellwood:2002}.
72: Transient spirals with a wide range of pattern speeds are present throughout the
73: evolution of the disk. Different pattern speeds result in spatially distinct locations of corotation resonances, allowing a star to undergo a series of resonant encounters in its
74: lifetime, riding the spiral waves across large portions of the galaxy while retaining a
75: mostly circular orbit. Stars born in-situ may therefore remain kinematically
76: indistinguishable from those that have come across the galaxy.
77:
78: In \citet{Roskar:2008} we presented first results from our $N$-body Smooth
79: Particle Hydrodynamics (SPH) simulations of disk galaxy formation in which
80: stars migrated radially due to their resonant interactions with transient spirals.
81: These migrations yielded radial stellar population gradient predictions,
82: which have recently been observed in resolved-star \citep{de-Jong:2007}, and
83: integrated-light studies \citep{Azzollini:2008, Bakos:2008}.
84: In conjunction with our models, these observational data
85: strongly imply that radial migration is an important process in
86: observed galactic disks. In this Letter, we present further analysis of the simulation
87: presented in \citet{Roskar:2008}, focusing on the wide-ranging implications of radial
88: migrations for the observable properties of stellar populations.
89:
90: \section{Simulation}
91:
92: The initial conditions are created as in \citet{Kaufmann:2006th}, and are designed to
93: mimic the quiescent stage following a last major merger when thin disk
94: formation commences. The system consists of a spherical dark
95: matter NFW halo \citep{Navarro:1997aa} in which we embed a spherical
96: halo of gas with the same initial profile. The gas halo is
97: initially in hydrostatic equilibrium. The total mass of the system is $10^{12}$M$_{\odot}$,
98: analogous to a Milky Way-type spiral galaxy, with the baryons contributing
99: 10\% of the mass. We resolve the system with $10^6$ particles per component,
100: resulting in initial mass resolutions of $10^6$M$_{\odot}$ and $10^5$M$_{\odot}$
101: for dark matter and gas respectively. Star particles form with masses that are a fraction
102: of the gas particle mass, resulting in typical star masses
103: around 3$\times$10$^4$M$_{\odot}$. To form a rotationally-supported
104: disk, we also impart angular momentum to the gas component
105: corresponding to a spin parameter value $\lambda$=0.039, as motivated by
106: cosmological $N$-body experiments \citep[e.g.][]{Bullock:2001}. We evolve the system
107: using the $N$-body + Smooth Particle Hydrodynamics code
108: \textsc{gasoline} \citep{Wadsley:2004mb} for 10 Gyr.
109:
110: Once the simulation begins, the gas cools and
111: collapses to the center of the halo, forming a thin rotating disk from the inside-out.
112: When the gas reaches densities and temperatures which allow for star formation,
113: the star formation and supernova (SN) feedback cycles are initiated
114: \citep{stinson:2006aa}. Since our disks form
115: without any \textit{a-priori} assumptions about the
116: interstellar medium (ISM) or the stellar populations present
117: in the disk, we can self-consistently follow the evolution of their properties as the disk grows.
118: Transient spirals cause radial redistribution of
119: stellar material in a manner analogous to the mechanism presented in SB02.
120: Due to space limitations, a detailed discussion of these processes is deferred until
121: a future paper.
122:
123: Our simulations do not account for the effects of evolution in a full cosmological context, but
124: in the standard $\Lambda$CDM paradigm mergers are significantly more important at early
125: epochs. The simulation presented here is therefore designed to mimic the
126: quiescent phase of disk galaxy formation during which the thin disk forms.
127: The increased resolution gained by stepping outside of the cosmological context allows us
128: to isolate the effects of key dynamical processes, which determine multiple
129: observational characteristics. Despite these simplifications, physical properties of our
130: model such as the rotation curve, star formation rate, disk scale-length,
131: and disk mass fraction, are analogous to those of observed systems. Pre-enrichment
132: of merger fragments would not change the details of
133: disk build-up, but simply offset the metallicity distribution.
134:
135: \section{Radial Migration and its Implications}
136:
137: Models of galactic chemical evolution have been enormously successful in explaining
138: the properties of stars in our solar neighborhood \citep{Matteucci:1989,Carigi:1996,
139: Chiappini:1997,Boissier:1999}. One example of such model
140: results is the age-metallicity relationship (AMR). The AMR is expected to arise
141: due to progressive enrichment of the ISM through stellar feedback.
142: Stars of the same age in the same
143: general region of the galaxy are therefore expected to have similar metallicities.
144: Indeed, early determinations of the AMR confirmed that
145: the mean trend of stars in the solar neighborhood is toward lower metallicity with increasing age
146: \citep{Twarog:1980}. Models, which assume that stars remain where
147: they are born and return their nucleosynthetic yields to their local ISM,
148: typically successfully reproduce this trend. However, evidence suggests that
149: a large amount of scatter is present in the AMR of
150: field stars \citep{Edvardsson:1993,Nordstrom:2004} and open clusters \citep{Friel:2002},
151: meaning that stars of the same age
152: are observed to have a wide range of metallicities.
153: In the framework of a galactic disk where
154: radial annuli are self-enriched, homogeneous, and isolated from one another, old
155: metal-rich stars or very young metal-poor stars are an impossibility.
156:
157: Allowing stars from slightly different regions of the disk to enter the
158: local sample due to the eccentricity of stellar
159: orbits has been considered as a possible explanation, but can
160: account only for up to 50\% of the observed scatter \citep{Nordstrom:2004,Binney:2007}.
161: The large amount of scatter in the AMR therefore poses
162: an important challenge for models of disk formation and should be considered
163: as large a constraint as the mean trend \citep{Carraro:1998}.
164: The unexpected degree of scatter implies that stars in the
165: solar neighborhood were either formed from a highly inhomogeneous ISM or have come
166: from wildly different parts of the galaxy.
167:
168: In our simulation, the latter option offers an enticing solution. In the left-most panel of
169: Figure~\ref{fig:solar_neighborhood} we show the distribution of birth
170: radii for stars, which at the end of the simulation are on nearly circular orbits within an
171: annulus analogous to the solar radius. We define
172: `solar radius' as a general region of the disk interior to the disk break, but
173: approximately 2-3 scalelengths from the center. In our simulated galaxy this region is
174: between 7-9 kpc (indicated by dashed lines in
175: Figure~\ref{fig:solar_neighborhood}). The black line represents
176: all stars while the blue and red lines show the distributions of metal poor
177: ([Fe/H] $< -0.3$) and metal rich ([Fe/H] $> -0.1$) stars respectively.
178: Roughly 50\% of all ``solar neighborhood'' stars
179: have come from elsewhere, primarily from the disk interior.
180: Interestingly, some metal poor stars have been scattered into the solar
181: neighborhood from the outer part of the disk. Such migration has recently been inferred from
182: observational data \citep{Haywood:2008}. Metal-rich stars, like our Sun, could have
183: originated almost anywhere in the Galaxy.
184:
185:
186: \subsection{Age-Metallicity Relationship}
187:
188:
189: %
190: % Age-Metallicity relation and G-dwarf figure
191: %
192:
193:
194:
195: \begin{figure*}
196: \centering
197: \plotone{f1}
198: \caption{Properties of stars in the solar neighborhood.
199: {\bf Left:} Histogram of birth radii for stars that end up in the solar neighborhood
200: on nearly circular orbits. The black, red, and blue lines represent all, metal-rich,
201: and metal-poor stars respectively.
202: {\bf Middle:} The age-metallicity relation: color
203: contours represent relative particle densities where point density is high. Diamonds and
204: error bars indicate mean values and dispersion respectively. Squares show
205: the AMR if stars are assumed to remain in-situ. A small horizontal
206: offset is applied to the two sets of symbols for clarity.
207: {\bf Right:} Metallicity distribution function (MDF): the simulated distribution
208: is shown with the solid black histogram; diamonds and asterisks show data from
209: \cite{Rocha-pinto:1996} and \cite{Holmberg:2007} respectively. The dashed
210: histogram is the MDF if stars are assumed to remain in-situ.
211: }
212: \label{fig:solar_neighborhood}
213: \end{figure*}
214:
215: It has previously been suggested that such large radial migrations
216: could account for the scatter in the observed
217: AMR (SB02, \citealt{Binney:2007}), but our models allow us to
218: examine the implications of migrations in a self-consistent simulation
219: of a growing and star-forming galactic disk.
220: We show the AMR as determined from our simulation in the middle
221: panel of Figure~\ref{fig:solar_neighborhood}. Colors indicate relative particle densities for
222: clarity in high-density regions. Diamonds and error bars indicate the means and standard
223: deviations in several broad age bins. With the exception of the oldest age bin, the AMR
224: is fairly flat. The relatively high number of very old ($> 9$ Gyr) extremely metal-poor
225: stars is partly a remnant of our initial configuration, which includes only
226: pristine, metal-free gas. This is likely to be different in a cosmological setting where
227: a fraction of the gas is accreted from subhaloes that have already formed stars
228: and been enriched by stellar feedback. Also, note that we assume each star is born with the
229: mean metallicity of the ISM at its birth radius so that we can more easily isolate the effects of
230: radial migration on the various observational relations.
231:
232: To illustrate the impact of the dynamical
233: evolution of the disk, we plot the AMR in these same age bins by using only stars that
234: actually formed in the solar neighborhood. This latter relation is analogous to that derived
235: by models which assume a dynamically non-evolving disk, essentially only depending on
236: the nucleosynthetic yields and star formation and gas infall rates at the solar circle.
237: In comparison to the full simulation,
238: the in-situ relation shows a much tighter
239: correlation between metallicity and age. Inclusion of full dynamical modeling
240: significantly flattens the AMR and increases its dispersion by at least a factor of two.
241: We note that although our simulations are not
242: designed to reproduce Milky Way observables, they successfully reproduce the
243: major qualitative features of the observations: the observed AMR is indeed relatively flat with a
244: high degree of scatter \citep{Edvardsson:1993,Nordstrom:2004},
245: which gradually increases with age \citep{Haywood:2008}.
246: This agreement strongly suggests that a large amount of radial migration
247: is the missing piece of the AMR puzzle.
248:
249: \subsection{Metallicity Distribution Function}
250:
251: If radial migration is important for the AMR, then it should also leave a significant imprint
252: on the metallicity distribution function (MDF), another key observational constraint for
253: models of Galactic chemical evolution. The simple, closed-box
254: chemical evolution picture where no material is allowed
255: to enter or leave a given radial annulus has long been recognized as inadequate
256: at explaining the local MDF
257: \citep{van-den-Bergh:1962}.
258: However, the dearth in the relative number of low-metallicity stars,
259: known as the G-dwarf problem, is usually explained by allowing for the inflow of metal-poor gas
260: \citep{Larson:1974,Lynden-Bell:1975}. Gas infall ``solves'' the G-dwarf problem
261: because it allows for prolonged star formation, which results in a relatively high number
262: of metal-rich stars.
263:
264: Our simulated MDF is shown with diamonds in
265: the right panel of Figure~\ref{fig:solar_neighborhood} compared to observational data
266: from \citet{Rocha-pinto:1996} and \citet{Holmberg:2007}, represented by
267: diamonds and asterisks respectively.\footnote{Note that to shift the model quantities
268: into the metallicity range relevant for the Milky Way, we apply an offset of +0.2
269: dex to the simulated metallicities. We show the observed distributions not to
270: claim quantitative agreement, but simply to illustrate the ramifications of
271: dynamical effects.}
272: We also plot the MDF of stars formed only in the solar neighborhood (shown with squares).
273: Compared to the observed MDF, the in-situ distribution is much narrower.
274: Radial migration introduces metal-poor and metal-rich stars into the
275: solar neighborhood from other parts of the disk, thereby broadening the MDF.
276:
277: Unlike the solar neighborhood AMR, which is highly dispersed and therefore difficult
278: to use as a robust guide for Galactic chemical evolution models, the MDF represents
279: a much more stringent constraint. It is clear from Figure~\ref{fig:solar_neighborhood}
280: that radial migration may significantly alter the nature of this fundamental property
281: of the solar neighborhood.
282:
283:
284: \subsection{Metallicity Gradients}
285:
286: %
287: % metallicity figure
288: %
289:
290: \begin{figure}
291: \centering
292: \plotone{f2}
293: \caption{
294: {\bf Left:}
295: Azimuthally averaged metallicity profile at 10 Gyr for stars in different age bins (young,
296: intermediate age, and old stars shown with solid, dashed, and dotted lines respectively).
297: {\bf Right:}
298: Azimuthally averaged metallicity profile of the gas at different times in the simulation.
299: }
300: \label{fig:metallicity}
301: \end{figure}
302:
303:
304: A third major constraint on models of disk formation and evolution is the radial
305: metallicity gradient.
306: If stars are assumed to remain in-situ, then stellar tracers of
307: different ages should provide information about the ISM metallicity from
308: which they spawned.
309: In the left-hand panel of Figure~\ref{fig:metallicity} we show the metallicity gradient as
310: determined from stellar tracers of different ages. The slope of the gradient
311: decreases with increasing age of stellar population. In most observational
312: studies, such a trend would be interpreted to mean that the slope of the ISM metallicity was shallower
313: in the past \citep{Maciel:2005}. In the right-hand panel of Figure~\ref{fig:metallicity}
314: we show the actual metallicity gradient in the ISM at different times in the
315: simulation. Rather than being flatter in the past, as implied by the
316: flat gradient of old stars, the ISM gradient was actually steeper and
317: evolved in the opposite sense to that indicated by the stellar tracers. Radial migration
318: causes more mixing in older populations, creating the appearance of flatter gradients
319: at early times and leading to a decoupling of stellar properties from their birth ISM.
320:
321: \subsection{Star Formation History}
322:
323: %
324: % SFH figure
325: %
326:
327: \begin{figure}
328: \centering
329: \plotone{f3}
330: \caption{
331: Star formation history in several broad radial bins.
332: }
333: \label{fig:sfh}
334: \end{figure}
335:
336:
337: Implications of radial migration can be extended beyond our own Galaxy to
338: extragalactic studies. In recent years, state-of-the-art
339: observations of resolved stellar populations have begun to probe
340: detailed, spatially-resolved star formation histories in disks of external
341: galaxies \citep[e.g.][]{Barker:2007a, de-Jong:2007, Williams:2008, Gogarten:2008}.
342: Typically, observed properties of stars in a given region of a
343: disk are used to construct a color-magnitude
344: diagram, which in turn is used to constrain stellar population
345: synthesis models to yield a best-fit star formation history and therefore
346: provide information about the formation of the disk \citep{Dolphin:2002}.
347: Again, the underlying assumption is that stars observed today are found
348: close to the radius of their respective birthplaces.
349:
350: In Figure~\ref{fig:sfh} we show the `migrated' (solid) and `un-migrated' (dashed) star
351: formation history (SFH) in our simulation for several broad radial bins.
352: The `migrated' SFH is constructed by considering the distribution of ages
353: in the desired region of the disk at the final snapshot of the simulation, much like an observer would
354: do given a snapshot of a present day galactic disk. The `un-migrated' SFH is
355: the actual SFH, recovered by
356: considering the times of formation for all stars formed within a given radial bin, regardless
357: of the final radius to which they eventually migrated. We show here all stars regardless
358: of the eccentricity of their orbits, but our conclusions remain the same even if we restrict our
359: analysis only to stars on nearly circular orbits.
360:
361: The discrepancy between the `migrated' and
362: `un-migrated' SFH is large, especially for the outermost regions of the disk. For example, the `solar
363: radius' bin (7 - 9 kpc) could erroneously be deduced to have a flat SFR for the past $\sim 8$ Gyr,
364: when it actually didn't reach current SFR levels until just $\sim 5$ Gyr ago.
365: Migration must therefore be accounted for even in extragalactic studies of stellar populations.
366:
367: \section{Conclusions}
368:
369: We have shown that radial migration caused by resonant interactions of stars with transient
370: spirals has significant repercussions on the observed properties of a wide range of stellar
371: population systems. The range of implications and possible solutions to outstanding problems
372: given by radial migration is appealing. Further work is required to characterize fully the
373: degrees and timescales of migration in spiral galaxies of different physical characteristics. On
374: the other hand, future observational efforts such as RAVE, GAIA and LSST will
375: provide intricate data sets that should further constrain the radial migration process in the
376: Milky Way. Ongoing and future HST projects such as
377: the ANGST and GHOSTS surveys, will require radial migration models in order to
378: reconstruct the SFH from resolved-star observations of nearby spiral disks and thereby
379: learn about the formation history of our galactic neighborhood.
380:
381: \acknowledgments We would like to thank T. Kaufmann for the use of his initial
382: conditions code. R.R. thanks C. Brook, J. Dalcanton and B. Gibson for useful discussions.
383: This research was supported in part by the NSF
384: through TeraGrid resources provided by TACC
385: and PSC. R. R. and T. R. Q. were supported by
386: the NSF ITR grant PHY-0205413 at the University of Washington.
387: V. P. D. was supported by an RCUK Fellowship at the University of Central Lancashire.
388: R. R. acknowledges support for a visit to the
389: University of Central Lancashire from a Livesey Award Grant held by V. P. D.
390:
391:
392: \bibliographystyle{apj}
393:
394: %\bibliography{/Users/rokstar/Documents/disks}
395: \begin{thebibliography}{40}
396: \expandafter\ifx\csname natexlab\endcsname\relax\def\natexlab#1{#1}\fi
397:
398: \bibitem[{{Azzollini} {et~al.}(2008){Azzollini}, {Trujillo}, \&
399: {Beckman}}]{Azzollini:2008}
400: {Azzollini}, R., {Trujillo}, I., \& {Beckman}, J.~E. 2008, \apjl, 679, L69
401:
402: \bibitem[{{Bakos} {et~al.}(2008){Bakos}, {Trujillo}, \& {Pohlen}}]{Bakos:2008}
403: {Bakos}, J., {Trujillo}, I., \& {Pohlen}, M. 2008, ArXiv e-prints, 807
404:
405: \bibitem[{{Barker} {et~al.}(2007){Barker}, {Sarajedini}, {Geisler}, {Harding},
406: \& {Schommer}}]{Barker:2007a}
407: {Barker}, M.~K., {Sarajedini}, A., {Geisler}, D., {Harding}, P., \& {Schommer},
408: R. 2007, \aj, 133, 1138
409:
410: \bibitem[{{Binney}(2007)}]{Binney:2007}
411: {Binney}, J. 2007, {Dynamics of Disks} (Island Universes - Structure and
412: Evolution of Disk Galaxies), 67--+
413:
414: \bibitem[{{Boissier} \& {Prantzos}(1999)}]{Boissier:1999}
415: {Boissier}, S. \& {Prantzos}, N. 1999, \mnras, 307, 857
416:
417: \bibitem[{{Brook} {et~al.}(2004){Brook}, {Kawata}, {Gibson}, \&
418: {Freeman}}]{brook:2004}
419: {Brook}, C.~B., {Kawata}, D., {Gibson}, B.~K., \& {Freeman}, K.~C. 2004, \apj,
420: 612, 894
421:
422: \bibitem[{{Bullock} {et~al.}(2001){Bullock}, {Dekel}, {Kolatt}, {Kravtsov},
423: {Klypin}, {Porciani}, \& {Primack}}]{Bullock:2001}
424: {Bullock}, J.~S., {Dekel}, A., {Kolatt}, T.~S., {Kravtsov}, A.~V., {Klypin},
425: A.~A., {Porciani}, C., \& {Primack}, J.~R. 2001, \apj, 555, 240
426:
427: \bibitem[{{Carigi}(1996)}]{Carigi:1996}
428: {Carigi}, L. 1996, RevMexAA, 32, 179
429:
430: \bibitem[{{Carraro} {et~al.}(1998){Carraro}, {Ng}, \&
431: {Portinari}}]{Carraro:1998}
432: {Carraro}, G., {Ng}, Y.~K., \& {Portinari}, L. 1998, \mnras, 296, 1045
433:
434: \bibitem[{{Chiappini} {et~al.}(1997){Chiappini}, {Matteucci}, \&
435: {Gratton}}]{Chiappini:1997}
436: {Chiappini}, C., {Matteucci}, F., \& {Gratton}, R. 1997, \apj, 477, 765
437:
438: \bibitem[{{de Jong} {et~al.}(2007){de Jong}, {Seth}, {Radburn-Smith}, {Bell},
439: {Brown}, {Bullock}, {Courteau}, {Dalcanton}, {Ferguson}, {Goudfrooij},
440: {Holfeltz}, {Holwerda}, {Purcell}, {Sick}, \& {Zucker}}]{de-Jong:2007}
441: {de Jong et~al.} 2007, \apjl, 667, L49
442:
443: \bibitem[{{Dolphin}(2002)}]{Dolphin:2002}
444: {Dolphin}, A.~E. 2002, \mnras, 332, 91
445:
446: \bibitem[{{Edvardsson} {et~al.}(1993){Edvardsson}, {Andersen}, {Gustafsson},
447: {Lambert}, {Nissen}, \& {Tomkin}}]{Edvardsson:1993}
448: {Edvardsson}, B., {Andersen}, J., {Gustafsson}, B., {Lambert}, D.~L., {Nissen},
449: P.~E., \& {Tomkin}, J. 1993, \aap, 275, 101
450:
451: \bibitem[{{Friel} {et~al.}(2002){Friel}, {Janes}, {Tavarez}, {Scott},
452: {Katsanis}, {Lotz}, {Hong}, \& {Miller}}]{Friel:2002}
453: {Friel et al.} 2002, \aj, 124, 2693
454:
455: \bibitem[{{Gogarten} {et~al.}(2008)}]{Gogarten:2008}
456: {Gogarten et~al.} 2008, \apj, submitted
457:
458: \bibitem[{{Haywood}(2008)}]{Haywood:2008}
459: {Haywood}, M. 2008, ArXiv e-prints, 805
460:
461: \bibitem[{{Holmberg} {et~al.}(2007){Holmberg}, {Nordstr{\"o}m}, \&
462: {Andersen}}]{Holmberg:2007}
463: {Holmberg}, J., {Nordstr{\"o}m}, B., \& {Andersen}, J. 2007, \aap, 475, 519
464:
465: \bibitem[{{Jenkins} \& {Binney}(1990)}]{Jenkins:1990}
466: {Jenkins}, A. \& {Binney}, J. 1990, \mnras, 245, 305
467:
468: \bibitem[{{Kaufmann} {et~al.}(2006){Kaufmann}, {Mayer}, {Wadsley}, {Stadel}, \&
469: {Moore}}]{Kaufmann:2006th}
470: {Kaufmann}, T., {Mayer}, L., {Wadsley}, J., {Stadel}, J., \& {Moore}, B. 2006,
471: \mnras, 370, 1612
472:
473: \bibitem[{{Larson}(1974)}]{Larson:1974}
474: {Larson}, R.~B. 1974, \mnras, 166, 585
475:
476: \bibitem[{{Larson}(1976)}]{Larson:1976}
477: ---. 1976, \mnras, 176, 31
478:
479: \bibitem[{{Lynden-Bell}(1975)}]{Lynden-Bell:1975}
480: {Lynden-Bell}, D. 1975, Vistas in Astronomy, 19, 299
481:
482: \bibitem[{{Maciel} {et~al.}(2005){Maciel}, {Lago}, \& {Costa}}]{Maciel:2005}
483: {Maciel}, W.~J., {Lago}, L.~G., \& {Costa}, R.~D.~D. 2005, \aap, 433, 127
484:
485: \bibitem[{{Matteucci} \& {Francois}(1989)}]{Matteucci:1989}
486: {Matteucci}, F. \& {Francois}, P. 1989, \mnras, 239, 885
487:
488: \bibitem[{{Mu{\~n}oz-Mateos} {et~al.}(2007){Mu{\~n}oz-Mateos}, {Gil de Paz},
489: {Boissier}, {Zamorano}, {Jarrett}, {Gallego}, \&
490: {Madore}}]{Munoz-Mateos:2007}
491: {Mu{\~n}oz-Mateos}, J.~C., {Gil de Paz}, A., {Boissier}, S., {Zamorano}, J.,
492: {Jarrett}, T., {Gallego}, J., \& {Madore}, B.~F. 2007, \apj, 658, 1006
493:
494: \bibitem[{{Navarro} {et~al.}(1997){Navarro}, {Frenk}, \&
495: {White}}]{Navarro:1997aa}
496: {Navarro}, J.~F., {Frenk}, C.~S., \& {White}, S.~D.~M. 1997, \apj, 490, 493
497:
498: \bibitem[{{Nordstr{\"o}m} {et~al.}(2004){Nordstr{\"o}m}, {Mayor}, {Andersen},
499: {Holmberg}, {Pont}, {J{\o}rgensen}, {Olsen}, {Udry}, \&
500: {Mowlavi}}]{Nordstrom:2004}
501: {Nordstr{\"o}m et al.} 2004, \aap, 418, 989
502:
503: \bibitem[{{Robertson} {et~al.}(2006){Robertson}, {Bullock}, {Cox}, {Di Matteo},
504: {Hernquist}, {Springel}, \& {Yoshida}}]{Robertson:2006}
505: {Robertson}, B., {Bullock}, J.~S., {Cox}, T.~J., {Di Matteo}, T., {Hernquist},
506: L., {Springel}, V., \& {Yoshida}, N. 2006, \apj, 645, 986
507:
508: \bibitem[{{Rocha-Pinto} \& {Maciel}(1996)}]{Rocha-pinto:1996}
509: {Rocha-Pinto}, H.~J. \& {Maciel}, W.~J. 1996, \mnras, 279, 447
510:
511: \bibitem[{{Ro{\v s}kar} {et~al.}(2008){Ro{\v s}kar}, {Debattista}, {Stinson},
512: {Quinn}, {Kaufmann}, \& {Wadsley}}]{Roskar:2008}
513: {Ro{\v s}kar}, R., {Debattista}, V.~P., {Stinson}, G.~S., {Quinn}, T.~R.,
514: {Kaufmann}, T., \& {Wadsley}, J. 2008, \apjl, 675, L65
515:
516: \bibitem[{{Sellwood} \& {Binney}(2002)}]{Sellwood:2002}
517: {Sellwood}, J.~A. \& {Binney}, J.~J. 2002, \mnras, 336, 785
518:
519: \bibitem[{{Stinson} {et~al.}(2006){Stinson}, {Seth}, {Katz}, {Wadsley},
520: {Governato}, \& {Quinn}}]{stinson:2006aa}
521: {Stinson}, G., {Seth}, A., {Katz}, N., {Wadsley}, J., {Governato}, F., \&
522: {Quinn}, T. 2006, \mnras, 373, 1074
523:
524: \bibitem[{{Tinsley}(1975)}]{Tinsley:1975}
525: {Tinsley}, B.~M. 1975, \apj, 197, 159
526:
527: \bibitem[{{Twarog}(1980)}]{Twarog:1980}
528: {Twarog}, B.~A. 1980, \apj, 242, 242
529:
530: \bibitem[{{van den Bergh}(1962)}]{van-den-Bergh:1962}
531: {van den Bergh}, S. 1962, \aj, 67, 486
532:
533: \bibitem[{{Wadsley} {et~al.}(2004){Wadsley}, {Stadel}, \&
534: {Quinn}}]{Wadsley:2004mb}
535: {Wadsley}, J.~W., {Stadel}, J., \& {Quinn}, T. 2004, New Astronomy, 9, 137
536:
537: \bibitem[{{White} \& {Frenk}(1991)}]{white:1991}
538: {White}, S.~D.~M. \& {Frenk}, C.~S. 1991, \apj, 379, 52
539:
540: \bibitem[{{Williams} {et~al.}(2008)}]{Williams:2008}
541: {Williams et~al.} 2008, \aj, submitted
542:
543: \end{thebibliography}
544:
545:
546:
547: \clearpage
548:
549:
550:
551: \end{document}
552: