1: % Gerald Cecil & James Rose

2: % To appear in Reports on Progress in Physics

3: %

4: \documentclass[10pt]{iopart}

5: \usepackage[latin1]{inputenc}

6: \usepackage{pslatex}

7: \usepackage{float}

8: \usepackage{graphicx}

9: \usepackage{amssymb}

10: \IfFileExists{url.sty}{\usepackage{url}}

11:                       {\newcommand{\url}{\texttt}}

12: \usepackage[]{natbib}

13: \bibpunct[; ]{(}{)}{,}{a}{}{;}

14: \makeatletter

15: 

16: \providecommand{\boldsymbol}[1]{\rm{\boldmath $#1$}}

17: \providecommand{\tabularnewline}{\\}

18: 

19: \usepackage{bm}

20: \eqnobysec

21: \newcommand{\apj}{\textit{Astrophys.\ J.\ }}

22: \newcommand{\aap}{\textit{Astron.\ Astrophys.\ }}

23: \newcommand{\nat}{\textit{Nature}\ }

24: \newcommand{\mnras}{\textit{Mon.\ Not.\ Roy.\ Ast.\ Soc.}\ }

25: \newcommand{\aj}{\textit{Astron.\ J.\ }}

26: \newcommand{\araa}{\textit{Ann.\ Rev.\ Astron.\ Astrophys.\ }}

27: \newcommand{\apjs}{\textit{Astrophys. J. Supp.\ }}

28: \newcommand{\pasp}{\textit{Publ. Astron. Soc. Pac.\ }}

29: \newcommand{\kms}{km~s$^{-1}$}

30: \newcommand{\Ms}{M$_\odot$}

31: 

32: \makeatother

33: \begin{document}

34: \review{Constraints on Galaxy Structure and Evolution from the Light

35: of Nearby Systems}

36: 

37: \author{Gerald Cecil and James A Rose}

38: \address{

39: Dept.\ of Physics \& Astronomy, University of North Carolina at Chapel

40: Hill, NC 27599-3255 USA}

41: \eads{\mailto{cecil@physics.unc.edu}, \mailto{jim@physics.unc.edu}}

42: 

43: \begin{abstract}

44: We review knowledge of galaxy structures obtained by their emitted light and

45: in the local universe where they can be studied in great detail.

46: We discuss the shapes of, and stellar motions within, galaxies, compositional

47: clues derived from their spectra, and what luminous matter implies

48: about their dark matter content. Implications on the

49: current theory of hierarchical galaxy formation are explored.

50: \end{abstract}

51: %\submitto{RPP}

52: 

53: \section{Introduction}

54: 

55: We have entered the age of `precision cosmology'.  Careful 

56: measurements of how objects of known radiant power have dimmed with distance,

57: the angular size of known distances at early epochs, and

58: the power spectrum of the matter distribution at the present epoch, have led

59: to a remarkably accurate cosmological model.  That is, the expansion history of the 

60: universe, which is governed by the density of visible and dark matter \citep{Freeman06} and by

61: dark energy, has been well constrained.  The tools of observational

62: cosmology are amazingly diverse.  Type Ia supernovae (SNe) --- the

63: deflagration/detonation of a white dwarf star after it accretes sufficient mass to exceed the

64: pressure support of degenerate electrons ---

65: provides a high-luminosity `standard candle' for

66: mapping the dimming of sources over cosmological distances, constraining

67: directly how the Hubble expansion parameter changes with cosmic

68: epoch.  Detailed maps of the power spectrum in angular scales of

69: temperature variations of the cosmic microwave background radiation reveal

70: various harmonics of density fluctuations at the time when the universe had cooled

71: sufficiently for electrons and protons to `recombine' into hydrogen atoms.

72: The scale of these fluctuations depends on the matter energy density

73: in the early universe, and the

74: angular scale observed at the present epoch depends on the expansion 

75: history since this recombination.

76: Hence, the temperature maps constrain the cosmological

77: world model.  Finally, the power spectrum of the 3D distribution of matter on

78: large scales, as traced by luminous galaxies, also constrains

79: both the spectrum of the initial density

80: fluctuations and the combination of matter and dark energy that governs

81: their evolution.

82: 

83: The recent success of observational cosmology poses both new challenges

84: and opportunities. Perhaps most exciting is to quantify structure growth.

85: In the favoured dark energy plus cold, dark matter ($\Lambda$-CDM)

86: model, non-baryonic dark matter (DM) is non-relativistic long before the

87: epoch where neutral hydrogen atoms (HI in astronomical jargon) first

88: form. From the minimum (Jeans) mass that can collapse by self gravity

89: at that time ($\gtrsim10^{6}$ solar mass, \Ms), cold dark matter

90: (CDM) clumps in a hierarchy of increasing mass.

91: 

92: This scenario has been simulated extensively into the non-linear regime, but

93: is largely untested by observations. Temperature anisotropies

94: in the cosmic microwave background radiation

95: probe only the \textit{start} of the linear regime of structure

96: formation as gravity drags on cosmic expansion. While anisotropies

97: support the idea that the seeds of non-linear structure have

98: Gaussian amplitude spectrum with random phases (i.e.,

99: are spatially uncorrelated), unclear is how coherent structure evolves

100: into the non-linear regime, and if the radiant power (luminosity

101: in astrophysical jargon) and mass scales and morphologies of the observed

102: structures are consistent with the CDM hierarchy.

103: 

104: To document how structure evolves from when galaxies first appeared

105: at redshift $z\sim5$ until the present epoch it is essential to understand

106: the evolution of visible baryons as density-biased tracers of the

107: underlying DM. Even studies that map DM more directly through gravitational

108: lensing must know the structure of the lens by its visible material.

109: Thus, all studies of structure evolution depend on a better understanding

110: of galaxies. 

111: 

112: Determination of our cosmological world model offers an exciting opportunity

113: for students of galaxy evolution. Prior interpretation of, for example,

114: number counts of galaxies as a function of redshift, depended on both

115: galaxy evolution and an uncertain world model. An accurate

116: model now decouples geometry from galaxy evolution observed

117: over a significant redshift, hence lookback time.

118: 

119: Advanced ground- and space-based telescope and detector technologies

120: from x-ray to millimetre wavelengths allow study of galaxy

121: evolution \textit{in situ} at high redshift. To mention only a few examples over the

122: past decade: the development of `optical' telescopes of up to 10.5-meter

123: aperture that are coupled to multi-slit/fibre spectrographs, advanced detectors

124: at infrared (IR) wavelengths, the advent of imaging at sub-millimetre

125: wavelengths, high-spatial resolution UV to near-IR images from

126: the Hubble Space Telescope (HST), detailed studies by

127: x-ray observatories of hot plasma in the Milky Way galaxy (MWg) interstellar

128: medium (ISM) and in the intracluster/group medium, and mid-IR views

129: from the Spitzer Space Telescope.

130: 

131: While there has thus been a stampede toward objects at high redshift,

132: to understand these challenging, photon-starved measurements require

133: multi-waveband constraints on comparable objects

134: nearby. We must assess how baryons have concentrated into galaxies

135: through, among other processes, star formation (SF) and evolution,

136: nucleosynthesis of heavy elements, and dynamical processing of stars

137: and dissipational ISM. Because galaxies are proximate enough

138: to interact strongly by tides, we must determine

139: how these processes are affected by environment (rich galaxy

140: cluster versus small group).

141: 

142: For those oriented toward cosmology, understanding galaxy evolution

143: is a necessary prerequisite to using galaxies in bulk to trace the

144: evolution of the dominant DM and to constrain the mysterious

145: dark energy. Others will research the diverse and fascinating physics

146: of galaxy evolution. Our review is aimed at those who are

147: preparing to follow either path. We assess what properties of galaxies

148: have been established recently in the \textit{nearby} universe

149: $(z\lesssim0.1,$ an insignificant lookback time compared to

150: the age of the universe). In number, 80\% are gas-rich spiral galaxies

151: (Sgs) and smaller dwarf irregular galaxies (dIrr's), the rest are

152: gas-poor elliptical (Eg) and S0 (called collectively, early-type)

153: galaxies and dwarf spheroidal galaxies (dSph's). We address questions

154: about galaxy structure and evolution using images and spectra of much

155: higher intrinsic spatial resolution and signal-to-noise ratio (SNR)

156: than can be obtained on high redshift counterparts. Our theme throughout

157: is the perhaps surprising degree of uncertainty that prevails over

158: most topics in galaxy structure and evolution.

159: 

160: Section \ref{sec:What-Can-We} introduces modern observations

161: of stars and DM. 

162: At the top level, we seek the main driver(s) of

163: galaxy evolution. Does a single process dominate? Our main observables

164: are the details within the bimodal separation of galaxies into luminous

165: `red and dead' early-type spheroids and blue star-forming discs.

166: How are these connected? Are Egs a single family differing only in

167: mass? Section \ref{sec:Spheroidal-Galaxies} examines their 3D structure

168: for universal characteristics. A consistent interpretation of their

169: surface photometry and kinematics requires the introduction of many dynamical

170: assumptions. Section \ref{sub:Dark-matter-content} discusses their

171: uncertain DM content. Section \ref{sec:Bulges-of-Spiral} examines

172: the apparently similar spheroidal bulges of Sgs, including those

173: of our MWg and our neighbour Sg Andromeda (M31). Are they primordial,

174: were they built from mergers,

175: or did they arise from secular (i.e., long-term) processes driven by instabilities?

176: 

177: In section \ref{sec:Thick-&-Thin} we study Sg discs of various thickness.

178: Are discs stable and axisymmetric? DM halos affect stability, so

179: we address secular effects in section \ref{sec:Matter-Transfer-Within}

180: and consider radial matter transfer. Are discs distinguished

181: only by their level of SF? We consider the multi-component ISM of

182: gas, dusty soot, magnetic field, and cosmic rays. At the bottom of

183: the potential well is often a supermassive black hole, best

184: studied by its induced stellar motions.

185: 

186: In section \ref{sec:Interactions-with-Other} we consider interactions

187: of stars with the hot gas and with other galaxies in rich clusters.

188: Do galaxies switch between spiral and ellipsoidal forms? How has environment

189: modified SF? We preview some new instrument capabilities in section 8, and

190: summarize in the last section. We assume that the

191: reader grasps the basics of the current cosmological world model 

192: \citep[for example the textbook by][]{Ryden02},

193: but we minimize astronomical jargon and suggest review articles as

194: broader topics emerge.

195: \ref{sec:What-(Astro-)Physical-Processes}

196: reviews key astrophysical processes to understand before considering

197: galaxy evolution; Table A.1 lists acronyms used throughout.

198: 

199: 

200: \section{\label{sec:What-Can-We}What do we learn from starlight?}

201: 

202: We first summarize observables from galaxy starlight.

203: Limiting uncertainties are \textit{systematic} (external), not statistical

204: (internal), in most astronomical datasets of relevance to galaxy

205: evolution because most dynamical parameters are not observed directly.

206: Instead, we shall see that they arise from correlations between, and

207: calibrations from, often heterogeneous photometric and kinematical

208: datasets whose inter-transformations introduce systematics. Uncertain

209: zero points and incompleteness biases that weigh points in a correlation

210: can be further confounded by the inadvertent inclusion into the study

211: sample of unrelated less luminous foreground or more luminous background

212: galaxies; the large range of galaxy luminosities and the complications

213: of patchy obscuration by dust mean that reliable distances are

214: a prerequisite to minimising such bias. Differential measurements

215: can minimize systematics. Computer simulations to quantify errors

216: are necessary in observational astrophysics, complicated by both the large and 

217: very small sizes of astronomical datasets.

218: 

219: 

220: \subsection{The Milky Way galaxy}

221: 

222: Observing stars over a wide range of masses and evolutionary states, 

223: one seeks to correlate their dynamical, SF, and chemical enrichment

224: histories with 3D space velocities, ages, and chemical compositions.

225: Stars live mostly in the relatively low luminosity phase of core hydrogen-to-helium

226: fusion. They spend the last $\sim$20\% of their lives in a series

227: of higher luminosity phases. High-mass stars have relatively brief lives

228: compared to low mass ones.

229: 

230: Only in the MWg can we obtain such extensive data on individual, especially

231: lower mass unevolved, stars. But only a minuscule fraction of its

232: $\sim10^{11}$ stars can be so detailed from our vantage point $8.0\pm0.4$

233: kpc \citep{Eisenhauer03} off-centre. 

234: Indeed, we can approximate

235: this program only within the $\sim50$ pc radius Solar Neighbourhood

236: where we can obtain accurate stellar distances by trigonometric

237: parallax from opposing points on the Earth's orbit. Otherwise, use

238: of secondary measures of distance built upon parallaxes amplifies

239: uncertainties. Over the last decade, parallaxes and photometry

240: from the Hipparcos satellite above atmospheric blur

241: has been combined with ground-based spectra and photometry

242: to calibrate the parameters of many nearby stars accurately \citep{Lebreton01}; future

243: astrometric space missions aim to expand this sphere a hundred-fold

244: in radius. Note that many stars with excellent parallax measurements

245: have relatively poor spectroscopy and even photometry; these deficiencies

246: are being rectified by ongoing surveys. 

247: Finally, our view of the MWg

248: in visible and UV light --- where the most powerful diagnostics

249: lie --- is obscured by interstellar dust. Only outside this waveband

250: does our view clarify along many sightlines.

251: 

252: Now consider velocity vectors. With parallax in hand and a radial

253: (line of sight, l.o.s.) velocity measured from the Doppler shift of

254: narrow absorption lines in the stellar photosphere, we must obtain

255: the star's tangential motion across the sky by measuring

256: at two widely spaced times

257: its sky position relative to more distant objects that define a fundamental

258: reference frame. As detailed

259: in \ref{sub:Gravitational-stellar-dynamics}, the orbit

260: inferred from the gravitational potential is the essential dynamical

261: input in the study of galaxy evolution. Key too are age

262: and chemical composition, which are extracted from spectral analysis

263: as summarized in \ref{sub:Stellar-Evolution}.

264: 

265: 

266: \subsection{Local Group galaxies}

267: 

268: The MWg helps to bind the Local Group of galaxies.

269: Across $\sim2.5$ million parsecs (Mpc), two large Sgs --- the MWg

270: and larger M31 --- dominate $\sim25$ satellites; a third smaller Sg

271: M33 ($\sim10\%$ of the mass of the MWg) is associated with M31. Together, the three Sgs emit 90\% of

272: the light of the Local Group. With M33, the other two large satellites of M31 are

273: the dwarf Eg M32 and dwarf S0 NGC 205. The largest satellites of the MWg are the barred

274: Large and irregular Small Magellanic Clouds (LMC and SMC). Local Group members

275: are still being discovered, especially near the dust obscured plane of the MWg.

276: 

277: Local Group members are close enough for some resolution into individual stars,

278: especially with HST. 

279: Luminous stars can be studied individually,

280: tending to be either young and massive or in

281: later evolutionary stages. 

282: The motions of the Magellanic Clouds across the sky

283: have been measured from stars \citep{Kallivayalil06a,Kallivayalil06b}.

284: The near side of the LMC is $10\%$

285: closer than the far. The resulting $20\%$ brightness difference of

286: stars of the same stellar type gives a rudimentary sense of the LMC's

287: 3D structure. All other Local Group members are either too small or distant

288: to yield reliable information along the l.o.s. Hence the MWg, and

289: to a far lesser extent the LMC, are the only galaxies for which 3D

290: information is obtainable directly. 

291: 

292: High-resolution spectra, from which chemical abundances can be derived,

293: have been obtained for individual stars in some MWg satellites.

294: Abundances constrain the chemical history of a galaxy.

295: For example, in section \ref{sub:The-lowest-luminosity} we compare

296: abundances in the Sagittarius dIrr

297: to those in the thick disc and halo of the MWg, to constrain the assembly

298: of the MWg.

299: 

300: 

301: \subsection{Other {}`nearby' galaxies}

302: 

303: Individual stars outside the Local Group are virtually unmeasurable.

304: Instead, the 3D structure of a galaxy is extracted from

305: starlight projected onto the sky and integrated spatially across each

306: image pixel. 

307: One must

308: rely on the integrated l.o.s.\ velocity field (velocity function) that is broadened

309: by dispersed stellar velocities.

310: In addition to the pixel's spectral content and mean light flux,

311: variations in the surface brightness (SB) that arise

312: from Poisson fluctuations in the number of bright, evolved stars within

313: each pixel can be used to constrain the galaxy distance \citep{Tonry88}.

314: 

315: Most galaxies have only a cosmic redshift-implied distance; only 

316: $\sim10,000$ have distances determined independently with accuracy 

317: sufficient to separate a galaxy's l.o.s. \textit{peculiar velocity} 

318: due to gravity from the bulk cosmic expansion redshift.

319: In short, we know at best only three of the six spatio-kinematical

320: descriptors of galaxy structure. In situations of high

321: symmetry, we can invert to constrain the spatial variation throughout

322: some of the volume. As discussed in section \ref{sub:Internal-motions},

323: asymmetries in absorption line profiles can constrain the structure of

324: less symmetrical systems.

325: 

326: 

327: \subsection{\label{sub:State-of-the}State of the art}

328: 

329: At its average SB $(V=22.5)$, the moonless night sky over angular

330: area 1 arc-second$^{2}$ emits the radiant flux of a giant star at

331: 0.2 Mpc or a Sun-like dwarf star at 32 kpc. Adaptive optics systems

332: on telescopes reduce the area of the effective

333: detection box, hence sky flux, at least another ten-fold, increasing

334: the detection distance of a single star at constant

335: SNR more than three-fold. From the ground, the SNR increases

336: at best as the square-root of the increased exposure time. HST

337: goes considerably deeper than ground-based facilities somewhat because of a

338: $\sim17\%$ on average reduced V-band background (which is still variable because of

339: unpredictable, cloud dependent Earthshine) but mostly because its diffraction

340: limited image tightens the detection box to $\sim0.02(\lambda/500\,\rm{nm})^{2}$

341: arc-second$^{2}$.

342: 

343: The radial light profiles of spheroidal galaxies

344: projected onto the plane of the sky

345: are characterized by an `effective radius' $R_{\rm{e}}$ within

346: which half of the galaxy luminosity lies. Figure \ref{fig:SB-of-NGC}

347: compares the SB of the moonless night-sky to those of spheroidal galaxies.

348: With diligent sky-subtraction, accurate photometry can be obtained

349: beyond $10R_{\rm{e}}$. Slit spectra can be obtained to $5R_{\rm{e}}$,

350: and micro-lens- or fibre-coupled spectrographs (field segmentation

351: techniques whose optics tend to scatter more light than open slits) can be

352: effective out to $1-2R_{\rm{e}}$. Until recently, selection effects

353: excluded galaxies of considerable angular extent but of constantly

354: low SB; section \ref{sub:Low-surface-brightness} discusses this population.

355: 

356: \begin{figure}

357: \begin{centering}\includegraphics[scale=0.37]{figure1.pdf}\par\end{centering}

358: \caption{\label{fig:SB-of-NGC}SB of the nearby Eg NGC 3379 (photometry from

359: \citealt{Capaccioli90}) with \fullcircle at radii $xR_{e}$

360: for $x=(1,5,10)$, the darkest night sky at a Chilean site (top \dotted)

361: and its 1\% level (lower \dotted), and the value at near-Earth orbit

362: (\broken). Because SB depends empirically on radius as the $1/4$ power,

363: this variable is plotted on the ordinate to highlight deviations; profiles

364: along both the apparent major and the minor axis are shown.

365: The galaxy profile within $\sim1.5$

366: arcsec radius is flattened by the blurring effects of the Earth's

367: atmosphere.}

368: \end{figure}

369: 

370: A complication, becoming historical, to transforming insights between

371: the local and distant universe is that we usually observe different

372: rest wavelengths at high redshift than we do nearby. Only recently

373: have UV-optimized spacecraft (for example, the Galaxy Evolution Explorer {[}\textit{GALEX}]

374: and Far Ultraviolet Spectral Explorer {[}\textit{FUSE}]) and near-IR ground-

375: and space-based spectrometers provided images and spectra with sufficient

376: wavelength coverage to transform accurately. Another complication

377: is the inadvertent inclusion of luminous background objects, often

378: star burst and/or active galactic nuclei (AGN), whose bright regions

379: are uncharacteristic of the underlying stellar mass. These ubiquitous contaminants

380: increase with survey volume and complicate the connection

381: to nearby objects.

382: 

383: Signatures of chemical abundances concentrate at wavelengths

384: shorter than $\lambda$550 nm \citep[see Figure 2 of][for example]{Bland-Hawthorn04},

385: the U, B, and V filter bands. One needs $\sim\lambda0.01$ nm resolution

386: $(R=\lambda/\Delta\lambda\sim40,000$) to measure accurately many elements mostly by minimising

387: systematic errors from otherwise blended lines. Stellar chemical abundances

388: are determined by high-dispersion spectroscopy of many, sometimes

389: weak absorption lines from key elements at various stages of ionization.

390: This program is feasible on even halo MWg stars with the largest

391: telescopes and most efficient spectrographs; an error of $\lambda10^{-4}$

392: nm in the equivalent width (defined in Figure \ref{fig:equiv})

393: of an absorption line yields an abundance error of 0.02 dex. F-K stars

394: emit most of their energy in the U- to L- ($\lambda$5000 nm) bands

395: that are all accessible from the ground. A SNR $>10$ measures {[}Fe/H]

396: with errors of $\pm0.1$ dex and errors in radial velocities of $\pm2$ \kms.

397: \footnote{A chemical abundance in brackets is measured in units of dex,

398: i.e. $[A/X]\equiv\log_{10}(N_A/N_X)-\log_{10}(N_A/N_X)_{\odot}$,

399: where $N_A$ and $N_X$ are the number density abundances

400: of elements $A$ and $X$. Positive dex means more metals than in the Sun.}

401: A six hour exposure with the latest spectrometers on the Keck telescopes measures the

402: intensities of H$\beta$ absorption lines in even dwarf Virgo cluster galaxies ($\sim16$ Mpc)

403: to $\sim\lambda0.02$ nm equivalent width

404: accuracy. In section \ref{sub:Evolutionary-Synthesis-Models} we show that

405: this translates into an age uncertainty of $\sim4$ Gyr for a 12 Gyr-old

406: star cluster; uncertainties are exponentially smaller for younger clusters.

407: 

408: \begin{figure}

409: \begin{centering}\includegraphics[scale=0.11]{figure2.pdf}\par\end{centering}

410: \caption{\label{fig:equiv}Defining equivalent width $w$ and residual intensity $r$

411: of an absorption line spectrum. The vertical axis is radiant flux,

412: zero at bottom. A ratio of residual intensities defines

413: an index. To define $w$, the area notched from the slanted stellar

414: continuum is repeated at left (or right). As shown, equivalent width depends

415: even in data of infinite SNR on how one draws `the' continuum

416: across a line and where the two sidebands are defined; this is especially

417: true at blue wavelengths where lines crowd. For an index based on

418: $r$, the continuum level is irrelevant but scattered light must be controlled.

419: $\sim45\%$ of Egs show

420: weak emission lines, which must be excised for a pure absorption

421: spectrum.}

422: \end{figure}

423: 

424: Over the past decade, several wide-field surveys were made of the MWg

425: and other galaxies. Notable in the MWg are microlensing

426: experiments (section \ref{sub:Evidence-in-the}) and the Wisconsin H$\alpha$

427: Mapper \citep[WHAM,][]{Hafner03} survey of the ionized ISM.

428: Extragalactic surveys include the 

429: the Sloan Digital Sky Survey \citep[SDSS,][]{York00}, the Two-Degree

430: Field Galaxy Survey \citep{Colless02}, and the

431: near-IR 2MASS \citep{Strutskie06} and

432: Deep Near Infrared of the Southern Sky surveys \citep[DENIS,][]{Epchtein94}

433: that dip closer to the plane of the MWg to find galaxies

434: that are mostly dust obscured at shorter wavelengths.

435: 

436: 

437: 

438: \section{\label{sec:Spheroidal-Galaxies}Ellipsoidal Systems}

439: 

440: Egs (and the bulges of disc galaxies), by their simple SB profiles and

441: lack of an extensive ISM and ongoing SF, would seem to be the most straightforward

442: stellar systems hence the natural starting point to study

443: galaxy evolution. To focus discussion, we first frame basic

444: questions within the context of the evolution of structure

445: in the universe:

446: 

447: \begin{enumerate}

448: \item Can the true 3D shapes of Egs and disc galaxy bulges be recovered

449: uniquely from their 2D SB? Are they primarily oblate, prolate, or

450: perhaps more complex (triaxial) spheroids?

451: \item Do Egs form a single homologous family (hence formation mode)

452: or fundamentally different families (formation modes)

453: over their entire million-fold mass range?

454: \item Did SF/chemical evolution complete in only

455: a few dynamical times ($<1$ Gyr), or is there evidence for extended

456: (perhaps discontinuous) SF?

457: \item Are Egs dominated by DM, and how is DM spread compared to luminous matter?

458: \end{enumerate}

459: We summarize work that has sought to understand SF

460: and chemical evolution histories in diverse environments and as functions of galaxy

461: luminosity (and mass) to test homology,

462: to establish the distribution of their intrinsic shapes, to learn

463: the dynamical r\^ole of a central supermassive black-hole, and to constrain their DM

464: content.

465: 

466: \subsection{Current observational capabilities}

467: 

468: Before addressing the observed nature of Egs, it is useful to establish

469: the capabilities of current large (6.5-meter+ aperture) ground-based

470: telescopes with efficient visible-band/near-IR detector mosaics and multi-aperture

471: spectrographs, and of rapidly expanding spectroscopic surveys. What

472: are the relative merits of large angular-area, high-SNR, spatially-resolved

473: data of a single (or a few) nearby object(s), as opposed to more modest

474: data on huge numbers of galaxies in surveys such as the SDSS? The

475: datasets deliver complementary insights, but it is instructive to

476: examine for specific questions whether exhaustive data on one object

477: are more illuminating than less extensive data on many. One can often

478: recover the distribution in properties of a large sample from statistical

479: analyses of the distribution of key observational parameters by imposing

480: Bayesian prior information. The biggest challenge for the large sample

481: approach are selection biases that occur when working with flux-limited

482: samples of spatially extended but diffuse objects. The problem is

483: more severe when comparing distant to nearby samples, where it is

484: generally uncertain how the luminous objects detected at high redshift

485: relate to the bulk of galaxies at the present epoch.

486: 

487: During the 1980's and 1990's, long-slit spectrographs on 4-meter aperture

488: telescopes measured Eg kinematics, statistics improving with increased

489: CCD sensitivities and pixels. As noted in section \ref{sub:State-of-the},

490: by $R_{\rm{e}}$ in a massive Eg, the SB of B-bandpass starlight

491: has dropped typically to that of the moonless night sky. By $2R_{\rm{e}}$,

492: one is working $\sim6$ times fainter than the sky SB. With a 4-meter

493: aperture telescope, one obtains spectra typically out to $\sim1.5R_{\rm{e}}$

494: in reasonable exposures. To constrain the internal structure of the

495: galaxy requires spectra at multiple position angles (PA's) on the

496: sky, at least along the photometric major and minor axes. In contrast, photometric

497: profiles for Egs could be traced down to 1\% of the sky SB with $\pm30\%$

498: systematic uncertainty from the removal of its time-variable OH emission-line

499: glow. Another limiting factor is often the care taken in acquiring

500: `flat field' images to calibrate spatial variations in photometric

501: sensitivity across the detector and through the optics. Particularly damaging is scattered

502: light within the instrument+telescope.

503: 

504: Three major advances have increased sensitivity substantially.

505: 

506: \begin{enumerate}

507: \item Spectroscopic studies have progressed from sparse, long-slits along

508: select PA's reaching $\sim2R_{\rm{e}}$ \citep[for example][]{Bak00}

509: to the full spatial coverage over $\sim R_{\rm{e}}$ of the SAURON

510: integral-field spectrograph (IFS) \citep{deZeeuw02} on $>70$ early-type galaxies in both clusters

511: and the field. SAURON constrains within $R_{\rm{e}}$, a complicated

512: region containing counter-rotating stars, warps, and other transient

513: structures. Its optics deliver $>1500$ contiguous spectra per exposure

514: in either a lower spatial resolution mode used to maximize spectral

515: SNR at large radii, or a higher resolution mode used on brighter

516: regions near the galaxy core. While mapping kinematics to 6 \kms\

517: accuracy, its spectra also yield fluxes of some spectral indices to

518: $\lambda$0.01 nm accuracy in equivalent width. We introduce SAURON results below,

519: as appropriate. Other smaller-field IFS's (some using an adaptive optics system to sharpen

520: ground-based images by nulling part of the effects of atmospheric

521: turbulence) have detailed stellar cores.

522: \item Development of the `nod-and-shuffle' technique of coordinated

523: telescope and detector shifts \citep{Glazebrook01} to subtract night-sky

524: OH emission spectral bands to Poisson statistical limits. Implemented

525: on 8-meter+ aperture telescopes with efficient spectrographs, this

526: technique cancels many systematic errors to obtain spectra at much

527: fainter light levels than previously. Note that long-slit spectroscopy

528: is an inefficient data collection tactic, because it covers outer

529: radii only sparsely where large areas are needed to boost SNR. With

530: an IFS, coverage is continuous over the field of view, with outer

531: regions receiving the higher weight necessary to boost SNR. Nevertheless,

532: because of the purity of sky subtraction through an open long-slit

533: plus nod-and-shuffle, the sparse coverage at very low SB of long-slit

534: spectra still complements the full coverage at higher SB of an IFS

535: whose field-segmenting optics scatter more light.

536: \item The advent of wide-angle imaging and spectroscopic surveys.

537: Specifically, the 2dF survey on the 3.9-meter Anglo-Australian

538: Telescope produced $\sim$250,000 galaxy redshifts, and the SDSS on

539: a 2.5-meter telescope produced multi-bandpass photometry of $>10^{7}$

540: galaxies and multi-fibre spectra of $\sim10^{6}$ galaxies. With proliferating

541: multiple-CCD cameras and dedicated robotic, wide-field telescopes

542: for executing all-hemisphere imaging and spectroscopy, the future

543: of large scale surveys is extremely promising.

544: \end{enumerate}

545: 

546: \subsection{\label{sub:Intrinsic-shape}Intrinsic 3D shape}

547: 

548: We begin discussion\footnote{Readers who are unfamiliar with gravitational

549: stellar dynamics should first read the material in Appendix A.2.}

550: of Egs with a deceptively simple question:

551: what is their intrinsic shape? \ref{sub:Building-a-mass} noted

552: that obtaining the 3D distribution from the Abel inversion of

553: the 2D SB distribution is ill-conditioned, being highly unstable to details of noise and

554: data artifacts; the inversion can only be achieved with an assumed

555: 3D symmetry. Thus, even very high quality SB data on the nearby, nearly

556: circularly symmetric `standard' Eg NGC 3379, yield no clear

557: answer on whether its slight ellipticity (Figure \ref{fig:SB-of-NGC})

558: implies oblateness, prolateness, or even triaxiality. It may well

559: harbour a weak, nearly face-on disc \citep{Capaccioli91}.

560: 

561: Eg shapes were therefore first constrained in galaxy samples by trending

562: the central SB against their degree of ellipticity.

563: galaxies of different ellipticity.

564: to map central SB versus degree of ellipticity. The seemingly well-posed

565: hypothesis that prolate galaxies viewed edge-on should have higher

566: SB than edge-on oblates has no clear-cut answer, although

567: the tendency has been to favour oblates \citep[for example][]{Marchant79}.

568: However, this statistical approach trades

569: the uncertainty of deprojecting a SB profile to 3D to that of extracting

570: the true parent distribution of 3D shapes from an observed distribution

571: of peak SB's and ellipticities, assuming of course that there \textit{is}

572: a well-defined parent distribution. In fact, \citet{Lambas92} conclude

573: that Egs are neither strictly oblate nor prolate.

574: 

575: Twisted isophotes whose ellipticity and PA of projected major axis change

576: with SB occur in many galaxies.

577: Section 4.2.3 of \citet{Binney98} shows that twisted isophotes

578: originate naturally when we view misaligned a triaxial system whose

579: principal axes are aligned and orthogonal at all intensities while

580: varying in axis ratio systematically with SB. 

581: But twists also result from strong tides between galaxies,

582: and indeed the most extreme ones are found in close galaxy pairs

583: \citep{diTullio79,Kormendy82}.

584: Add the aforementioned possible

585: disc lurking in some Egs, and 3D shapes cannot be determined.

586: 

587: In summary, a seemingly simple question about the supposedly

588: simplest galaxies leads to a surprisingly unsatisfactory result. To

589: progress, one must investigate kinematics.

590: 

591: 

592: \subsection{\label{sub:Internal-motions}Internal motions}

593: 

594: By the mid-1990's it was clear that most luminous Egs have isophotes

595: whose major axial PA's twist with radius, implying triaxial shapes.

596: Because triaxial shape implies a large fraction of stellar

597: orbits with additional isolating integral(s) (see Appendix A.2.3)

598: beyond $E$ and $L_{\rm{z}}$

599: \citep{Schwarzschild79}, the distribution function 

600: $f$ will have anisotropic velocity dispersions.

601: A signature of a rotating system with isophotes twisted by projected

602: triaxiality is rotation along \textit{both} projected axes, which

603: is seen in many Egs \citep[for example][]{Schechter79}.

604: Evidence for anisotropic velocity dispersions is more subtle. Any

605: such system will flatten more than could arise from the observed rotation.

606: This qualitative global statement lends itself naturally to quantification

607: by the tensor virial theorem (see Appendix A.2.5),

608: as developed by \citet{Binney78}, further discussed in

609: section 4.3 of BT, and reexamined by \citet{Binney05} in the light of

610: the new capabilities of IFS's.

611: 

612: The main indication of a triaxial shape is misaligned rotation, with

613: motion from tube orbits (Appendix A.2.4) streaming along the true minor-axis,

614: but observed along the \textit{projected} minor axis of an apparent

615: axisymmetric galaxy. 

616: Asymmetric absorption-line profiles are quantified with up to 6th-order Gauss-Hermite

617: velocity moments.

618: Combining absorption-line long-slit spectra to

619: map rotation speeds and surface photometry to map shapes, one interprets

620: with Jeans models (first two moments) from the tensor Virial Theorem or 

621: more general three-integral models (higher moments too)

622: to constrain the internal motions of spheroids. 

623: 

624: In an axisymmetric galaxy viewed edge-on that rotates about the z-axis,

625: the potential energy tensor \begin{equation}

626: W_{xx}=W_{yy};\,\, W_{ij}=0\,\,(i\neq j)\end{equation}

627: and similarly for tensors

628: $\boldsymbol{\Pi}$ and $\boldsymbol{T}$ (random and ordered motion); in addition,

629: $T_{zz}=0$. The tensor Virial Theorem then gives\begin{equation}

630: \frac{2T_{xx}+\Pi_{xx}}{\Pi_{zz}}=\frac{W_{xx}}{W_{zz}}\end{equation}

631: If we define \begin{equation}

632: 2T_{xx}=\frac{1}{2}\int\rho\overline{v_{\phi}}^{2}d^{3}x=\frac{1}{2}Mv_{0}^{2}\end{equation}

633: with $M$ the total mass, $v_{0}^{2}$ the mass-weighted mean-square

634: rotational part of the velocity function, and $\sigma_{0}^{2}$ the mass-weighted

635: mean square random part of the velocity function, we obtain\begin{equation}

636: \frac{v_{0}^{2}}{\sigma_{0}^{2}}=2(1-\delta)\frac{W_{xx}}{W_{zz}}-2\end{equation}

637: with $0\le\delta\le1$ measuring the anisotropy of the velocity dispersion

638: tensor (i.e., $\Pi_{zz}=(1-\delta)\Pi_{xx}$). If the isodensity surfaces

639: are ellipsoids, then $W_{xx}/W_{zz}$ depends only on their average

640: photometric ellipticity $\epsilon$ and not on their density profile

641: $\rho(r)$ along galaxy radii. If $v_{0}$ and $\sigma_{0}$ can also

642: be estimated by averaging spectra, then \full\ lines in Figure

643: \ref{fig:kormendy} plot the relation between the spatially averaged

644: $v_{0}/\sigma_{0}$, ellipticity $\epsilon$, and anisotropy

645: parameter $\delta$.

646: 

647: \begin{figure}[t]

648: \begin{centering}\includegraphics[scale=1.3]{figure3.pdf}\par\end{centering}

649: \caption{\label{fig:kormendy}How observables (ratio of maximum rotational

650: to mean random motions $V_{0}/\sigma_{0}$ and photometric ellipticity $\epsilon=1-$axial

651: ratio) averaged over an oblate spheroidal Eg are related for various

652: values of the velocity anisotropy $\delta$ \full. Lines

653: (\broken) map to the left the theoretical $\delta$ curves drawn at right

654: as our viewing angle is varied from edge-on (right) to

655: face on (lower left). Plotted are light-weighted averages across individual

656: luminous Egs ($\times$ from \citealp{Illingworth77,Bertola75})

657: that mostly show no rotational support, and Virgo dwarf Egs (\fullcircle\

658: from \citealt{Geha02} and \citealt{vanZee04}) and lower luminosity

659: Egs ($+$ from \citealp{Davies83}) that mostly do; grey denotes

660: an upper limit of the measured $V_{0}/\sigma_{0}$.}

661: \end{figure}

662: 

663: 

664: There are two complications. First, (3.4) assumed that we view

665: the galaxy edge-on to its true major axis. Otherwise the observed

666: eccentricity and velocity function alter. The dashed

667: lines in Figure \ref{fig:kormendy} show different $\delta$ hence

668: viewing angle $i$ that is usually unknown, so $\delta$ is degenerate.

669: However, different parts of the diagram indicate clear oblate rotation

670: or anisotropic flattening.

671: Second, although the tensor Virial Theorem is a global average of observables,

672: most of its results arise from a few long-slit spectra

673: through the nucleus along the major axis (and minor axis

674: and crossed at $45^{\circ}$ when more spectra are obtained).

675: Spectra normally do not extend beyond $R_{\rm{e}}$, so are not

676: global averages. An assumption must convert from observables

677: such as maximum rotational velocity and central velocity

678: dispersion to global values required by the tensor Virial Theorem. 

679: 

680: Despite these deficiencies, our understanding of Eg dynamics is based on

681: this Theorem.

682: Early work \citep{Illingworth77,Bertola75}

683: found that most luminous Egs flatten largely by velocity anisotropy,

684: not rotation. Subsequent work \citep{Davies83} led to the surprise

685: that lower luminosity dwarf Egs \textit{are} flattened mostly by

686: rotation. Recent attention has focused on Virgo cluster dwarfs

687: \citep{Geha02,vanZee04}. While most do not rotate,

688: some do so fast enough to flatten. Figure \ref{fig:kormendy}

689: summarizes these results.

690: 

691: 

692: The tensor Virial Theorem gives the global balance between the inferred rotation and

693: velocity dispersion anisotropy (hence further infers triaxiality and

694: the potential), at least when we have a reliable inclination.

695: Details on the internal structure of Egs can come

696: only by fitting to some dynamical model the spatially resolved spectra

697: of both streaming and random motions over an area of the galaxy. 

698: Statler and collaborators have fit spectra mostly from \citet{Davies88}

699: out to $\sim$1.4 $R_{\rm{e}}$ using Stäckel potential models with minimal

700: prior assumptions about dynamics. The resulting bivariate

701: probability distribution of flattening $c_{L}$ and degree of triaxiality

702: $T$ often yields an oblate best fit for an individual galaxy, but

703: the result is usually not well constrained. Instead, \citet{Bak00}

704: fit to the parent shape distribution of \citet{Davies88}'s 13 Egs,

705: using the same Bayesian models. They seek the shape distribution

706: of the entire family, which can be estimated more robustly than the

707: shape of any single galaxy. Figure \ref{fig:Statler} contours the intrinsic

708: shape of their sample as a bivariate likelihood distribution in $T$

709: and $c_{L}$. At left is the distribution from spectra and

710: surface photometry, with both prolates and oblates present but with

711: oblates more likely and strong triaxiality rare. In contrast, the

712: right panel shows the parent distribution if only photometry is used.

713: While flattening is still well constrained, triaxiality is not.

714: 

715: \begin{figure}[H]

716: \begin{centering}\includegraphics[scale=3.5]{figure4.pdf}\par\end{centering}

717: \caption{\label{fig:Statler}\citep[From][used with permission.]{Bak00} The parent distribution

718: of intrinsic shapes of a sample of 13 Egs with high-quality, multiple position-angle

719: slit spectra, plotted in the space of triaxiality $T$ and flattening

720: $c_{L}$ of the light distribution. Contours enclose 68\% and 95\%

721: probabilities under the assumption that the parent population has

722: no preferred orientation to our l.o.s. a) Using photometric \textit{and}

723: kinematic data, a bimodal distribution is evident, with more probability

724: of oblate than prolate objects. The probability of them having very

725: triaxial shape is low. b) Photometric data alone constrain only the flattening.}

726: \end{figure}

727: 

728: In summary, in proceeding from analyses of individual galaxies (e.g.,

729: NGC~3379) to a set drawn from a distribution, shapes are better constrained

730: but at the expense of added dynamical assumptions on the parent distribution.

731: \citet{Bak00} find that if rotation is assumed to be mainly `disc-like'

732: (decreasing rapidly from the major axis), then triaxiality is indeed

733: rare. However, if rotation is assumed to have `spheroid-like'

734: constant rotation speed at all latitudes, then triaxiality is common.

735: Results are still hostage to model assumptions.

736: 

737: The advent of IFS spectra that map the velocity function across the inner $\sim R_{\rm{e}}$

738: now allow application of the tensor Virial Theorem to a representative sample of Egs

739: with fewer assumptions \citep{Binney05}.

740: For example,

741: \citet{Verolme02} use SAURON IFS spectra to constrain

742: the shape of M32. Assuming axisymmetry, they obtain intrinsic flattening

743: $0.68\pm0.03$. \citet{Krajonovic05} is another illustration of modern orbit

744: fitting. \citet{Cappellari05}

745: is a preliminary analysis of part of the SAURON sample, in the spirit

746: of Figure \ref{fig:kormendy}. They show that those galaxies divide

747: into weakly triaxial and nearly isotropic slow-rotators, and nearly

748: oblate and anisotropic fast-rotators.

749: 

750: The reader may wonder if this discussion has led to a

751: result. Clearly, the CBE has not been solved uniquely.

752: That results (for example on whether a sample of Egs is

753: mainly prolate or oblate) are as varied as the underlying assumptions

754: shows that we are still far from uncovering shapes and establishing

755: the dynamics of Egs, especially using only photometry. Yet,

756: significant progress \textit{has} been made: the tensor Virial Theorem, supported by isophotal

757: twists evident in some Egs, shows conclusively that the more massive

758: Egs are flattened mainly by velocity dispersion anisotropy, while

759: intermediate mass Egs tend to be oblate rotators. This is a

760: key constraint on scenarios of galaxy evolution. Further progress

761: should come soon as IFS samples are analysed with statistical techniques

762: such as those of \citet{Bak00}; for example, see \citet{Statler04}.

763: 

764: 

765: \subsection{\label{sub:Global-scaling-parameters}Global scaling parameters}

766: 

767: Recall a question posed at the start of this Section: are Egs

768: a single homologous family over their full mass range?

769: While some investigators consider the range $\sim10^{6}$

770: to $\sim10^{12}$ \Ms, we adopt the conservative

771: approach that Eg luminosities range between $-18>M_{B}>-22$, dwarf Egs

772: range between $-13>M_{B}>-18$, and galaxies still fainter fall into

773: the dSph and/or globular cluster category. While the connection of dSph's

774: (studied extensively in the Local Group) to the main family of d/Egs is unclear,

775: we now ask: have all d/Egs had similar formation histories?

776: Does mass to light ratio ($\Upsilon$, in solar units) vary with luminosity? 

777: 

778: 

779: \subsubsection{Photometric properties.}

780: 

781: Consider first the azimuthally averaged radial SB profiles of Egs.

782: While the azimuthal average may blur important differences between

783: objects (e.g., the possibility that different kinds of Egs tend to

784: show `discy' versus `boxy' isophotes), the radial SB profile

785: is a natural starting point. Given the degree of uncertainty that

786: arose from questioning the 3D structure of Egs in section \ref{sec:Spheroidal-Galaxies},

787: one should not be surprised that there is controversy over 

788: whether Egs form a single homologous

789: family in their radial SB profiles $I(R)$.

790: Much early work emphasized the difference between Eg and dwarf Eg profiles.

791: Specifically, luminous (hence, massive) Egs are well fit by the \citet{deVau48}

792: $r^{1/4}$ law\begin{equation}

793: \ln(I(R)/I_{\rm{e}})=-7.67[(R/R_{\rm{e}})^{1/4}-1]\end{equation}

794: with $I_{\rm{e}}\equiv I(R_{\rm{e}})$

795: and factor 7.67 to ensure that half of the total intensity comes from

796: $R<R_{\rm{e}.}.$ On the other hand, dwarf Egs (and Sg discs) are fit better by the

797: exponential profile\begin{equation}

798: \ln(I(R)/I_{\rm{e}})=-R/R_{\rm{e}}\end{equation}

799: A successful exponential fit does not imply that all Egs are

800: actually disc dominated, although it does raise the question of whether

801: some dwarf Egs masquerade as S0 galaxies. However, a different radial

802: SB profile does indicate that d/Egs may differ fundamentally. Further

803: evidence comes by plotting $\mu_{0}\equiv I(0)$ the central SB against

804: $M_{\rm{B}}$, the logarithm of the blue luminosity of the galaxy.

805: Starting from low luminosities, $\mu_{0}$ increases monotonically

806: with luminosity; at the higher luminosity end the trend reverses:

807: $\mu_{0}$ decreases as luminosity increases \citep[for example][]{Kormendy85}.

808: When few Egs with $M_{\rm{B}}\sim-18$ had accurate surface photometry,

809: it was generally thought that both the SB profiles and $\mu_{0}$'s

810: change character in a sharp transition near luminosity $M_{\rm{B}}\sim-18$.

811: 

812: However, \citet{Caldwell83} showed that Egs form a continuous colour-magnitude

813: sequence, without breaking at $M_{\rm{B}}=-18$. In fact, recent

814: papers \citep[for example][]{Trujillo01,Graham03} have advocated

815: that Egs of all luminosities are fit by the more general \citet{Sersic68}

816: profile\begin{equation}

817: \ln(I(R)/I_{\rm{e}})=-b_{n}[(R/R_{\rm{e}})^{1/n}-1]\label{eq:Sersic}\end{equation}

818: with $b_{n}$ a function of the shape parameter $n$ set so that half

819: of the intensity is within $R<R_{\rm{e}}$ , and parameter $n$

820: varies gradually from 1 at the low luminosity end to $\ge4$ (the

821: de~Vaucoulers profile) at the high end. \citet{Graham03} find continuity

822: in the plot of $\mu_{0}$ (or $\mu_{\rm{e}}$) against $M_{\rm{B}}$:

823: Figure \ref{fig:Graham-plot.} shows that $\mu_{0}$ first

824: increases from low luminosity, then transitions smoothly near $M_{\rm{B}}\sim-20.5$

825: to decrease with further increasing luminosity. They show that a

826: S\'ersic profile fits very well throughout the galaxy for all lower

827: luminosity Egs; \citet{Graham05} presents the case that Egs form

828: a single family in their photometric profiles. Only at higher masses

829: must it be modified in the inner regions to avoid over-predicting

830: SB, and \citet{Graham03} argue that more massive Egs harbour a central

831: supermassive black-hole, which \citet{Valluri98} among others have shown may

832: make orbits near the centre chaotic, thereby lower the central

833: stellar mass density. However, \citet{Statler04} analyze one Eg studied

834: by SAURON and show this does not appear to have happened.

835: 

836: \begin{figure}

837: \begin{centering}\includegraphics[scale=0.3]{figure5.pdf}\par\end{centering}

838: \caption{\label{fig:Graham-plot.}From Figure 2 of \citet[used with permission]{Graham05},

839: showing the reversing trend in central blue SB at the horizontal \dotted\

840: for Egs more luminous than log galaxy

841: luminosity $M_{B}\sim-20.5$ in blue-magnitude units. The diagonal solid line (\full)

842: shows the correlation established from different data sets (different symbols, see paper) 

843: at low luminosity.}

844: \end{figure}

845: 

846: What to make of these results? On the one hand, enough free parameters

847: can always fit all galaxy profiles onto a single profile. On the other,

848: the evidence of Graham and collaborators does alleviate the

849: previously claimed discontinuity in Eg profiles from exponential to

850: $r^{1/4}$, and also explains plausibly the transition in central

851: SB as a function of galaxy luminosity. We therefore feel that, excluding

852: dSph's, there is no longer a strong case to be made from photometric

853: profiles that Egs divide into two distinct families.

854: 

855: 

856: 

857: \subsubsection{\label{sub:Kinematics:-the-fundamental}Kinematics: the fundamental

858: plane.}

859: 

860: Moving from photometry to kinematics, the tensor Virial Theorem has been applied to

861: other properties of early-type galaxies that correlate tightly over

862: at least $R_{\rm{e}}$, if not globally. The space defined by SB

863: $I_{\rm{e}}$ and velocity dispersion $\sigma$ is the surface\begin{equation}

864: R_{\rm{e}}\propto\sigma^{1.24\pm0.07}/I_{\rm{e}}^{0.82\pm0.02}\label{eq:FPlane}\end{equation}

865: \citep[for example][]{Jorgensen96} called the `fundamental plane'

866: (FP). The FP shows that the final radius of an Eg inside its DM halo

867: is set by the collapsing mass to within a factor of 2 over a hundredfold

868: range of mass. The narrow range of possible final configurations of

869: Egs is set by dissipation. A homologous collapse that ends in virial

870: equilibrium with constant $\Upsilon$ would yield $R_{\rm{e}}\propto\sigma^{2}/I_{\rm{e}}$.

871: Systematics of the observed deviation from (\ref{eq:FPlane}) through

872: breakdown of homology --- the so-called `tilt of the FP'

873: --- show the dependence on galaxy mass or other global parameters

874: of stellar $\Upsilon$ (hence the stellar population via the initial mass function, IMF) and

875: stellar and DM phase space densities. 

876: If the dynamical basis of these systematics can be understood, then their

877: trends with lookback time can provide insights into the mass build up of galaxies

878: and can test models of galaxy formation and evolution.

879: 

880: How these dependencies vary among galaxies sets the

881: intrinsic, thin scatter of the FP. For example, \citet{Jorgensen96}

882: find that\begin{equation}

883: \Upsilon_{\rm{fp}}\propto\sigma^{0.86}\label{eq:ML}\end{equation}

884: with 25\% scatter. 

885: (Such small scatter yields fairly accurate distances, enabling

886: study of Eg peculiar velocities.)

887: From extensive photometry by combining ground-based

888: and HST I-band images of 25 galaxies and SAURON spectra within

889: $R_{\rm{e}}$, \citet{Cappellari06} find that $\Upsilon_{\rm{fp}}$

890: can also be written within experimental uncertainties as a simple power law

891: of either luminosity or mass alone.  They derive mass densities

892: from a Jeans model and distribution function that depends only on $E$ and $L_{\rm{z}}$,

893: deprojecting their photometry and stellar velocities under an assumed

894: axisymmetry and fixed $\Upsilon$. 

895: They find only weak dependence of $\Upsilon$ on the

896: assumed inclination angle $i$ (good, because \citet{Krajonovic05} show that $i$ is

897: constrained only weakly by IFS spectra).

898: Next, fixing $i$ from the Jeans model, they use Schwarzschild's orbit

899: summing technique (\ref{sub:Stellar-orbits-and}) to construct

900: more realistic three-integral but still axisymmetric models that use more 

901: moments of the velocity function, not just the first and second used in the

902: Jeans model. Fits to spectra yield the same constant $\Upsilon$ as

903: from the Jeans model within scatter, demonstrating that the derived

904: $\Upsilon$ is independent of the Jeans assumption of a homologous

905: distribution function. The noisier $\Upsilon$ derived from their three-integral fits

906: to the velocity function yields the tight correlation \begin{equation}

907: \Upsilon(<R_{\rm{e}})=(3.8\pm0.14)(\frac{\sigma_{\rm{e}}}{200\,\rm{km s}^{-1}})^{0.84\pm0.07}\label{eq:MLsigma}\end{equation}

908: $\Upsilon$ is found to depend mainly on galaxy mass and then, at

909: fixed mass, on $\sigma_{\rm{e}}$, a trend related to galaxy compactness.

910: The scatter around (\ref{eq:MLsigma}) is small enough to exclude

911: significant triaxiality for at least the fast rotators in their sample, supporting

912: the assumptions that led to $\Upsilon$. Examining the scatter as

913: a function of luminosity, they find that the slow rotators, hence generally

914: higher mass Egs of triaxial shape (section \ref{sub:Internal-motions}),

915: have larger $\Upsilon$ than the fast rotators hence generally lower mass

916: oblate Egs. They show that non-homology is responsible for $\lesssim7\%$

917: of the variation in (\ref{eq:MLsigma}), most scatter coming instead

918: from intrinsic variations of $\Upsilon$ inside $R_{\rm{e}}$ around

919: typical stellar values that are consistent with simple stellar populations,

920: see section \ref{sub:Evolutionary-Synthesis-Models}) models.

921: They find that the `virial mass' defined as \begin{equation}

922: M{}_{\rm{virial}}=(5.0\pm0.1)R_{\rm{e}}\frac{\sigma_{\rm{e}}^{2}}{G}\end{equation}

923: \textit{is} a reliable, unbiased estimator of mass within $R_{\rm{e}}$.

924: This result confirms that

925: $R^{1/4}$ profiles and $\sigma_{\rm{e}}$ measured in a large aperture \textit{can}

926: measure mass build up even at high redshift.

927: 

928: 

929: \subsection{\label{sub:Dark-matter-content}Dark matter content}

930: 

931: \citet{Cappellari06} compare their dynamical estimate of $\Upsilon$

932: with that inferred from simple stellar population

933: synthesis to reproduce observed spectral

934: indices. They reject a \citet{Salpeter55} IMF because the \citet{Kroupa01}

935: form scatters less, and infer that the IMF barely varies among

936: galaxies. They conclude that $\sim30\%$ of the total mass

937: within the $R_{\rm{e}}$ of early-type galaxies is dark, consistent

938: with the few lensing cases \citep[for example][]{Ferreras05,Rusin05}.

939: Only a variable DM fraction within $R_{\rm{e}}$ can explain the

940: observed scatter.

941: 

942: \citet{Mateo98} reviews DM constraints from the internal kinematics

943: of Local Group dSph galaxies. Masses and central velocity dispersions are usually

944: derived using King models (Appendix A.2.2), which assume isotropic velocity dispersions

945: and that mass follows light. \citet{Kleyna01} obtain radial velocities

946: of many stars in the Draco dSph, find a flat/slowly rising velocity

947: dispersion at large radii, hence argue for an extended DM halo. As

948: a group, dSph's often have $\Upsilon\ge200$, making them nearly as

949: DM dominated as some low-SB galaxies (section \ref{sub:Low-surface-brightness}).

950: For example, \citet{Kleyna05} have studied 7 candidate stars in the

951: faint, 100 kpc distant, recently discovered \citep{Willman05} Ursa

952: Major dSph. They identify five likely members, assume an isotropic

953: velocity function with constant $\Upsilon$, and obtain a central

954: $\Upsilon=500$ solar from colours; they argue that this is a lower

955: limit to the global value because the galaxy is so dark. 

956: Is this system

957: a prototype of the `DM satellites' that linger around mature

958: galaxies at the end of $\Lambda$-CDM simulations \citep{Navarro97}?

959: \citet{Mateo98} obtained $\log\Upsilon=2.5+10^{7}/(L/\rm{L}_{\odot}),$

960: i.e., that each early-type dwarf has a luminous component with $\Upsilon=2.5$

961: solar that is embedded in a DM halo of $10^{7}$ \Ms. Central black-holes

962: more massive than $10^{4}$ \Ms\  are rejected in all systems except

963: Ursa Major \citep{Maccarone04} and Draco.

964: 

965: Further insights on DM have emerged from studies of planetary nebulae in Egs.

966: A planetary nebula converts up to 15\% of the white dwarf core's radiation to the $\lambda500$

967: nm {[}O~III] emission line, allowing velocities to be measured accurately

968: in early-type galaxies up to 30 Mpc distance on a 4-meter aperture

969: telescope. 

970: They are important probes in more massive spheroidal galaxies because their

971: emission lines are isolated easily from starlight with narrow-band

972: filters, and their kinematics can be mapped very efficiently using

973: a slitless spectrograph over a large field of view \citep{Douglas02}

974: and out to $5R_{\rm{e}}$ to make application of the tensor Virial Theorem more reliable.

975: 

976: However, they can be $<3$ Gyr old, a `young' population

977: in early-type galaxies. Thus their orbits have probably not relaxed

978: dynamically to the same state as their parent galaxy, for example

979: if their progenitor stars were formed in the late merger of a gas

980: rich system. \citet{Romanowsky03} have studied $\sim100$ planetary nebulae in each

981: of four lower-luminosity Egs. They found that velocities decline

982: by $\sim60\%$ in Keplerian fashion between one and $3R_{\rm{e}}$,

983: obviating need for DM. However, they note that such a decline could

984: also result from mostly radial stellar orbits, as might occur in an

985: Eg formed when two discs merge. They derive $\Upsilon=6.4\pm0.6$

986: solar that is consistent with their population synthesis, compared

987: to $20-40$ found in more luminous Egs. \citet{Napolitano05} obtain

988: similar results. These observations seem to leave little room for

989: DM within $2R_{\rm{e}}$ in lower luminosity Egs.

990: 

991: However, challenged by such compelling data, \citet{Dekel05} choose

992: more realistic stellar and DM density profiles, simulate disc galaxy

993: mergers, find that they can reproduce the observed planetary nebula velocity function with $\beta(r>R_{\rm{eff}})\equiv(1-\sigma_{\theta}^{2}/\sigma_{\rm{r}}^{2})\approx0.75$

994: and a significant DM halo, and establish that such values arise from

995: their simulations because radial orbits form preferentially in the

996: outer parts of the merger remnant whatever the gas content of the

997: progenitor galaxies. $\beta$ correlates with tidal strength, being

998: higher for more head-on collisions. It exceeds the $\le0.2$ predicted

999: for DM halos, suggesting that the planetary nebulae do not trace the DM. \citet{Merritt93}

1000: reached a similar conclusion, estimating that hundreds of nebular velocities

1001: would be needed to probe DM in the absence of other kinematical constraints.

1002: In fact, \citet{Pierce06} find that the velocity dispersion of globular clusters

1003: in NGC 3379 is instead constant with radius, arguing for a substantial

1004: DM halo. Clusters on initial radial orbits will soon disrupt tidally,

1005: reducing the orbital anisotropy of this population. 

1006: 

1007: These studies

1008: emphasize the need for multiple DM tracers to provide consistency

1009: and highlight the uncertainties from dynamically `young' populations.

1010: Using globular clusters and planetary nebulae separately is perilous.

1011: 

1012: How is DM distributed spatially? Equipotential surfaces are always

1013: more spherical than the matter distribution. It is therefore likely

1014: that a DM halo around a spheroidal galaxy is close to spherical. It

1015: may, however, be offset from the photometric centre of the galaxy,

1016: and in simulations is also found at small radii to be `cuspy'

1017: \citep[for example][]{Navarro97}. Kinematical evidence for a central

1018: cusp of DM is inconclusive because of uncertain stellar disc and bulge

1019: mass decompositions, the presence of distinct dynamical subsystems

1020: in the core, and plausible variations in $\Upsilon$, dust and stellar populations

1021: throughout the volume.

1022: 

1023: 

1024: \subsection{Stellar Content}

1025: 

1026: Having discussed the structure and dynamics of Egs, we turn to questions

1027: that bear on their formation and evolution.\footnote{A reader unfamiliar with the

1028: basics of stellar evolution should first review Appendix A.1 where some standard

1029: abbreviations are defined.} Unfortunately, despite

1030: growing capabilities to detect and study the luminous content of galaxies

1031: to lookback times of their formation, we cannot observe the assembly

1032: of the dominant DM halos. Instead, we must rely on observations of

1033: the SF and chemical evolution histories of the luminous baryons to infer how underlying

1034: DM evolves. Given compelling evidence from both observations \citep[for example][]{Searle73,Larson80}

1035: and numerical models \citep[for example][]{Mihos96} of galaxies undergoing

1036: gravitational interactions/mergers, it is generally agreed that an

1037: Eg that forms by merging anything other than two gas-free progenitors

1038: will show enhanced SF and chemical evolution. So, a key question is: to 

1039: what extent

1040: can SF/chemical evolution histories of galaxies be extracted from their spatially

1041: integrated starlight? In particular, how well can one determine the

1042: mean age and chemical composition of stars in an Eg? And, if SF indeed

1043: continued after a few dynamical times as the DM halo formed,

1044: how reliably can one measure the SF/chemical evolution history? Naturally, to measure

1045: discontinuities in this history would be particularly useful.

1046: 

1047: 

1048: \subsubsection{\label{sub:Evolutionary-Synthesis-Models}Key issues for integrated

1049: light spectroscopy.}

1050: 

1051: Before addressing results from integrated colours and spectra, we

1052: highlight some important complications. 

1053: 

1054: \begin{itemize}

1055: \item The evolution of a coeval stellar population is highly non-linear

1056: in time. That is, the integrated spectrum first evolves very rapidly,

1057: then slows as the MS turn-off (see Appendix A.1.5)

1058: drops to lower masses. In fact, the models in Figure \ref{fig:Spectral-evolution-of}a

1059: show that the population ages logarithmically.

1060: 

1061: \begin{figure}

1062: \begin{centering}\includegraphics[scale=0.8]{figure6.pdf}\par\end{centering}

1063: \caption{\label{fig:Spectral-evolution-of}(a) Evolution of the standard

1064: simple-population synthesis

1065: model of \citep[][used with permission]{Bruzual03} for solar metallicity, 

1066: using their GALAXEV models

1067: (\url{http://www2.iap.fr/users/charlot/bc2003/index.html}). Times since

1068: the end of the star burst are given in Gyr; note very small changes

1069: $4-13$ Gyr after the burst. Between $0.1-1$ Gyr, all Balmer absorption

1070: lines strengthen from late-B to early-F stars; this is the

1071: standard diagnostic of a recent star burst. Over the same interval,

1072: the `Balmer break' from the Balmer continuum limit evolves

1073: into the `4000 \AA~ break' from numerous prominent

1074: metal lines to the blue. In gray are the \textit{GALEX} FUV, \textit{GALEX}

1075: NUV, and SDSS filter bandpasses. Sometimes, ages are determined crudely from the

1076: relative intensities of light in adjacent bandpasses.

1077: (b) A series of simple-population synthesis 

1078: spectral models of increasing age and decreasing

1079: metallicity (top to bottom) show how the two parameters are nearly degenerate.

1080: However, close examination of the spectra reveals that the apparent degeneracy can

1081: indeed be broken with spectra that resolve individual line strengths.}

1082: \end{figure}

1083: 

1084: \item The integrated $\Upsilon$ of the evolving population also increases

1085: dramatically and logarithmically from very small values for a young

1086: population. Thus, even limited recent SF tends to overwhelm light

1087: from the old population, skewing sharply the luminosity-weighted mean

1088: age of a galaxy toward the age of the youngest stars. Such bias is,

1089: of course, useful if one wishes to detect even small amounts of recent

1090: SF from mergers.

1091: \item Degeneracy between age and metallicity was the most serious complication to 

1092: confront the first studies in integrated

1093: light. Specifically, Figure \ref{fig:Theoretical-distribution-of}

1094: shows that an older, metal-poorer population mimics a younger

1095: metal-richer one. However, as detailed below, measuring mainly

1096: age-sensitive hydrogen Balmer absorption lines coupled with metal features such

1097: as Fe and/or Mg lifts this degeneracy \citep{Worthey94}. Age/metallicity

1098: and dust reddening are a second degeneracy that alters photometric

1099: colours and the overall spectral energy distribution. Dust reddens

1100: starlight, making a galaxy look older and/or metal-richer. However,

1101: individual spectral features are unaffected.

1102: \item The most vexing hence interesting current issue for integrated light

1103: studies is that element abundance ratios in Egs do not track those

1104: in the Sun and Solar Neighbourhood stars. As we discuss below,

1105: recent work shows that the $\alpha-$element/Fe ratio exceeds solar

1106: in Egs, indicating more rapid chemical evolution than occurred locally.

1107: While this result provides rich insights on galaxy evolution, it

1108: unfortunately complicates models of integrated spectra.

1109: \item Extracting a galaxy age from an integrated spectrum demands far more

1110: of stellar evolution models than extracting the age of a resolved

1111: star cluster from its CMD. For a CMD age, stellar modellers match

1112: only the position and shape of the MSTO (Appendix A1.5). In contrast, an

1113: integrated spectrum is a composite of light from many parts of the HR diagram. So, the

1114: shape of isochrones must be correct across the HR diagram, not just

1115: at the MSTO. Even more challenging, isochrones must have the correct

1116: stellar \textit{number density} across the HR diagram. Equivalently,

1117: the timescales for \textit{all} advanced phases of stellar evolution

1118: must be correct. For example, \citet{Schiavon02a,Schiavon02b} show that discrepancies

1119: between the age of the MWg globular cluster 47 Tuc derived from its integrated light

1120: and from its CMD may arise largely because isochrones under-predict

1121: stars along the upper part of the RGB.

1122: \end{itemize}

1123: 

1124: \subsubsection{\label{sub:Results-for-the}Results for the dwarf elliptical galaxy

1125: M32.}

1126: 

1127: Given these complications, the reader should question whether reliable

1128: mean ages and chemical abundances of Egs can be extracted from integrated

1129: spectra and photometric colours. In fact, this issue has been controversial

1130: for decades. Pioneering models \citep{Spinrad71,Faber72} discovered

1131: that Eg integrated spectra are dominated by the light of metal-rich, mostly old

1132: stars. Moreover, there were strong indications that successful models

1133: require stars with CN and Mg features that are only reproduced by

1134: stars in the Solar Neighbourhood whose line strengths exceed greatly

1135: those found in stars of solar abundance, a result that is now seen

1136: in the context of enhanced-to-solar $\alpha$-element/Fe abundances.

1137: \citet{Faber72} found that the mean metallicity of Egs increases

1138: with mass. However, these studies could not constrain Eg ages beyond

1139: showing that Eg spectra are clearly dominated by older stars.

1140: 

1141: \citet{Oconnell80}, in another landmark study of galaxy spectra,

1142: analysed the integrated spectrum of the nucleus of the dwarf Eg M32 and

1143: concluded that the Balmer absorption lines are too strong to be explained

1144: just by an old, metal-rich population. Rather, M32 must contain MSTO

1145: stars of intermediate age. Clearly, that Egs can contain such stars

1146: has important implications on their formation and evolution, so O'Connell's

1147: controversial assertion has been much scrutinized and has made M32

1148: a `Rosetta Stone' for integrated light studies of galaxies.

1149: In fact, its nucleus is bright enough to obtain a very high SNR spectrum

1150: (SNR $>100$ per $\lambda$0.1 nm) in only a one-hour integration

1151: on a small, 1.5-meter aperture telescope. Moreover, due to its low

1152: ($\sim$80 \kms) central velocity dispersion, spectral features

1153: are better resolved than in the centres of massive Egs (central velocity

1154: dispersion $\sim$300 \kms).

1155: 

1156: Several ways have been proposed to enhance metal and Balmer absorption

1157: lines without resorting to a multi-age population \citep[see][for a summary]{Renzini86}.

1158: Specifically, populations with a range in metal abundance can reproduce

1159: the observed spectrum, i.e., an age-metallicity degeneracy argument.

1160: Alternatively, a few hot ($\sim$10,000 K) stars could

1161: strengthen Balmer absorption lines without resorting to a major intermediate

1162: age population; these might be blue stragglers --- upper MS stars from

1163: a very minor $\sim$1 Gyr old population --- or blue HB stars from a

1164: metal-poor population. Finally, perhaps non-solar abundance ratios

1165: have so contaminated the bandpasses of the indices used to measure

1166: key spectral features that we are not modelling properly the integrated

1167: spectrum.

1168: 

1169: Clarifying the situation of M32 and other Egs led to key

1170: advances. Tinsley and collaborators \citep[for example][]{Gunn81}

1171: introduced the now standard evolutionary population synthesis, i.e., comparing integrated

1172: spectra of galaxies to a grid of model populations based on isochrones

1173: and either empirical or synthetic spectra of stars.

1174: (Synthesis replaced optimized fits to a galaxy spectrum from an observed

1175: spectral library with few stellar astrophysics constraints.)

1176: Another advance was the development of the Lick system of

1177: spectral indices \citep[for example][]{Faber85}, a set of passbands

1178: defined to isolate and quantify important spectral features. In a key

1179: advance, \citet{Worthey94} modelled carefully

1180: the age-metallicity degeneracy. He showed that plotting a Balmer line

1181: index (e.g., strength of H$\beta$) against a metal-sensitive index

1182: (e.g., strength of the Fe I $\lambda$5270 feature) decouples the

1183: age-metallicity degeneracy. Finally, population synthesis models have

1184: begun to incorporate non-solar abundance ratios (NSAR) to model spectra

1185: and indices \citep{Maraston03}. This requires modelling the effects

1186: of NSAR on the absorption features due to the abundance changes

1187: on individual features \citep{Tripicco95}, and on the internal stellar

1188: structure and evolution due to changes in, for example, the interior

1189: opacity from changing proportions of major electron donors such C,

1190: N, and O.

1191: 

1192: The present situation is encouraging. Simple stellar population models now

1193: produce sensible ages and metallicities for well understood stellar

1194: systems such as globular \citep{Schiavon02a,Schiavon02b} and open clusters

1195: \citep{Schiavon04}. For M32, recent analyses

1196: agree that an intermediate age population is unavoidable to reproduce

1197: the integrated spectrum of its central $\sim$3 arc-seconds, both

1198: \citet{Worthey04} and \citet{Rose05} finding that the light-weighted

1199: mean age doubles from the central value ($\sim$3-4 Gyr) at

1200: the nucleus to $R_{\rm{e}}$. Mean metallicity halves, so these

1201: competing effects `conspire' to keep the integrated B-V colour

1202: roughly constant. \citet{Rose05} demonstrate with a spectral indicator

1203: sensitive to light from hot stats that the enhanced Balmer absorption

1204: lines (over that expected for a uniformly old population) cannot be

1205: attributed to a small population of hot stars. Hence, a more substantial

1206: intermediate age population is indeed necessary to explain the strong

1207: Balmer lines. Moreover, both \citet{Worthey04} and \citet{Rose05}

1208: find only modest departures from solar abundance ratios in M32, simplifying

1209: the interpretation of its integrated spectrum.

1210: 

1211: 

1212: \subsubsection{Results for large samples of elliptical galaxies.}

1213: 

1214: We turn to results from large surveys of spatially integrated Eg spectra.

1215: \citet{Trager00} studied 50 Egs in both the low-density field and

1216: clusters, and found a significant range of light-weighted ages. The

1217: range is widest for field Egs and for those with lower velocity dispersions

1218: $\sigma$. As well as being older on average, larger $\sigma$ Egs

1219: have larger NSAR, with {[}Mg/Fe] enhanced heavily in the most massive

1220: ones. \citet{Caldwell03} obtain similar results in a study of nearly

1221: 200 Egs that cover a wide range in $\sigma$ (hence luminosity and

1222: mass). Most striking is their result that low $\sigma$ galaxies scatter

1223: in age far more than those with high $\sigma$, i.e., low-mass Egs

1224: have had more prolonged SF histories. In contrast, in a study of early-type

1225: galaxies in the Fornax cluster, \citet{Kuntschner00} found that all

1226: Egs have similar old ages, but that metallicity increases with $\sigma$.

1227: In a recent study, \citet{Sanchez06} analyse integrated spectra of

1228: nearly 100 field and cluster Egs. They find that Egs in low-density

1229: environments range in age more than their clustered counterparts,

1230: with those in low-density environments tending to be more metal rich and younger

1231: by $\sim$1.5 Gyr. For these galaxies,

1232: an age-metallicity relation is found when metallicity is measured by Fe 

1233: (but not by Mg). While a well-defined

1234: mass-metallicity relation is found for clustered galaxies from both

1235: Fe and Mg, for field galaxies a mass-metallicity relation is only

1236: evident in Mg. They also find that lower luminosity galaxies spread

1237: more in age than their higher luminosity counterparts. 

1238: 

1239: Remarkably, these studies have uncovered little evidence for a mass-metallicity

1240: relation among field Egs. Most lower mass Egs have had prolonged

1241: SF histories and slower chemical evolution than their higher mass counterparts. In

1242: contrast, clustered Egs are systematically older than Egs in the

1243: field, hence much more consistent with the hypothesis that SF occurred

1244: rapidly and ended early to produce a well-defined mass-metallicity

1245: relationship.

1246: 

1247: How do these results relate to predictions of the $\Lambda$-CDM hierarchy?

1248: As mentioned, one cannot simply compare galaxy SF/chemical evolution histories

1249: to the simulated build up of CDM halos. Rather, simulations must incorporate

1250: parametrized simplifications of baryonic astrophysics to follow luminous particles.

1251: For example, \citet{Somerville99}

1252: have predicted SF histories of Egs, finding, unsurprisingly, that they follow

1253: the assembly timescale of DM halos. Specifically, at all

1254: masses (luminosities) one expects from the stochastic build-up of the hierarchy

1255: to produce a range in light-weighted mean ages. Too, more massive

1256: galaxies tend to have prolonged SF (hence younger light-weighted

1257: mean ages) than low mass galaxies. Moreover, at given mass, SF is

1258: prolonged for galaxies in low density field environments compared to

1259: in rich clusters, as predicted by \citet{Kauffman98}. 

1260: As \citet{Proctor02}

1261: and \citet{Sanchez06} discuss, it is encouraging news for $\Lambda$-CDM models

1262: that light-weighted

1263: mean ages of Egs spread significantly, and also that Egs in clusters

1264: are, in the mean, older than those elsewhere. However, that mean ages

1265: of lower mass Egs are younger (and less enhanced $\alpha$-elements)

1266: than for massive ones counters a straightforward

1267: prediction of hierarchical models.

1268: 

1269: Inverted observation/model trends clearly require

1270: a different model prescription for SF feedback, i.e.,

1271: something that quenches SF in the deeper potentials of more massive

1272: galaxies. It has been argued recently 

1273: \citep[for example][]{Benson03,Bower06} that the solution to this problem

1274: and to others in predicting the form of the galaxy luminosity function

1275: at both low and high redshift, is connected with the formation of

1276: supermassive black-holes in the centres of more massive galaxies and the subsequent

1277: quenching of SF (see section 6).

1278: 

1279: Finally, one must recognize that 

1280: existing data cover only the centres of Egs. Light-weighted

1281: mean ages and metallicities are dominated by the nuclear region in

1282: the spectrograph slit or few arc-second diameter fibre,

1283: generally only a small fraction of $R_{\rm{e}}$. Long-slit studies

1284: of M32 indicate that the intermediate age population concentrates

1285: centrally \citep{Worthey04,Rose05}. Dwarf Egs in the Virgo cluster, which

1286: typically have a younger light-weighted mean age \citep{Caldwell03}, 

1287: are bluer in the nucleus

1288: \citep{Vader88}, an `inverse' radial colour gradient. Together,

1289: these results argue that the young and intermediate age populations

1290: in Egs may be mostly nuclear. Unfortunately, little is known of population

1291: gradients in Egs, due partly to centrally concentrated light that

1292: makes extremely challenging the high SNR studies out to $R_{\rm{e}}$

1293: necessary to disentangle age from metallicity. Radial colour gradients

1294: traced to faint SB \citep[for example][]{Vader88,Peletier90,Wu05} reveal that

1295: massive Egs have redder centres. The straightforward interpretation

1296: is that their nuclei are more metal-rich. However, disentangling age

1297: from metallicity with only visible-band colours is problematic. With

1298: rapid progress in near-IR imaging, adding near-IR colours 

1299: increases discrimination between age and metallicity thirtyfold

1300: \citep{Cardiel03,MacArthur04}. (We

1301: caution that modelling near-IR spectra requires understanding all the

1302: issues of stellar mass loss and convection discussed in Appendix A.1.5;

1303: see \citealt{Lee06}.) \citet{Kobayashi99}

1304: summarized work on spectral index gradients in Egs from many sources.

1305: For massive Egs at least, 

1306: they concluded that universal NSAR's (as characterized

1307: by {[}Mg/Fe] and significant metallicity gradients) exist.

1308: 

1309: 

1310: \subsubsection{Stellar populations: the next steps.}

1311: 

1312: The bottom line from the previous section is that current data support

1313: hierarchical formation in some ways, and challenge it in others. What

1314: near-term, perhaps decisive, improvements in both observations and modelling of integrated

1315: spectra can we anticipate? There are two data improvements.

1316: First, results to date have solved the first-order

1317: problem of the light-weighted mean ages and chemical compositions

1318: of Egs. Now we must discriminate between multiple populations. Specifically,

1319: to what extent is the $3-4$ Gyr mean age of M32's nucleus a composite

1320: old and young population, or perhaps instead multi-episodic? Second, as mentioned

1321: above, data on radial population gradients to constrain scenarios

1322: of Eg assembly are sparse, especially for low mass Egs.

1323: 

1324: The first question of distinguishing mean ages and metallicities in

1325: one or more populations is addressed by using multiple spectral features

1326: that range widely in wavelength. A younger population, being bluer

1327: than the older, will contribute proportionately more blue light than

1328: red. Thus light-weighted mean ages from bluer spectral indices give

1329: younger ages than those from redder ones \citep[for example][]{Sanchez06,Schiavon04}.

1330: Such subtle distinctions require very high SNR and very reliable models

1331: of different spectral indices. For example, clear problems in modelling

1332: the H$\delta$ line as an age indicator come from `contamination'

1333: by a nearby CN molecular band that can only be modelled confidently

1334: by using the correct NSAR prescription \citep{Proc07}. So the problem

1335: is not strictly one of better quality data; we also need better models.

1336: 

1337: Much modelling effort is underway to construct

1338: fully consistent NSAR grids that incorporate not only the stellar

1339: interior/evolution effects of NSAR, but also effects on the atmospheric

1340: structure and its emergent spectrum. Ongoing work to synthesize

1341: spectra for various NSAR prescriptions will eventually replace empirical

1342: spectral libraries and their limitation to Solar Neighbourhood abundance

1343: ratios \citep{Coelho05}.

1344: 

1345: Two tactics yield reliable radial age and abundance gradients

1346: in galaxies. First, measure gradients out to interesting radii by going as

1347: faint as is possible reliably.

1348: This requires not only many

1349: photons hence large-aperture telescopes and high-throughput

1350: spectrographs and detectors, but also very accurate sky subtraction and

1351: control of scattered light.

1352: Second, cover an extended region with an IFS, do not view a narrow

1353: slice of it through a slit.

1354: 

1355: The second advance is the advent of vast spectral surveys such

1356: as the 2dF and the SDSS. Recent work on extracting the SF histories

1357: of many galaxies in the SDSS have not targeted Egs \citep{Jimenez05}

1358: specifically. In fact, fingering Egs for separate analysis is unwise

1359: because morphology is ephemeral in the $\Lambda$-CDM

1360: hierarchy. As discussed above, with exceptional SNR on a single galaxy

1361: across many different radii,

1362: one might distinguish multiple episodes of SF from a light-weighted

1363: mean age. That would be a true advance.

1364: A clever example of constraining SF by statistics

1365: is \citet{Trager00},

1366: who posit a `frosting' model for Egs to explain the distribution

1367: of light-weighted mean ages in their sample. Frosting adds one episode

1368: of young stars to an old population. Leverage increases

1369: when galaxy ages and metallicities from intermediate and

1370: high redshift surveys \citep[for example DEEP-2 and COMBO-17]{Faber05} are combined with

1371: current epoch large samples like the SDSS. Clearly, constraints strengthen

1372: when statistical properties of galaxies must be reproduced at

1373: multiple epochs. It is beyond our scope to review the growing observations of

1374: galaxy evolution at high redshift; we mention only

1375: that recent data prefer `quenched' SF (SF that ends

1376: abruptly rather than a single recent burst) over frosting

1377: \citep{Faber05}.

1378: 

1379: 

1380: \subsubsection{\label{sub:The-lowest-luminosity}The lowest luminosity (Local Group dwarf)

1381: spheroidal galaxies: resolved stellar populations.}

1382: 

1383: In testing models to form and evolve Egs, much is gained by studying

1384: the smallest and faintest galaxies because they are the most common.

1385: Hundreds of dwarf Egs have been catalogued in the nearby Virgo \citep{Bingelli85}

1386: and Fornax \citep{Ferguson89} clusters. To test the Eg model in the

1387: extreme, one can study individual stars in Local Group dSph's. dSph's are very

1388: low luminosity ($M_{\rm{V}}>-14$), low central SB ($\mu_{0,\rm{V}}>22$

1389: mag arcsec$^{2}$), gas-poor galaxies that barely rotate. Their relation

1390: to more massive Egs and to dwarf Egs (which are somewhat more luminous

1391: and have brighter centres than dSph's) is unclear. \citet{Mateo98}

1392: and \citet{Grebel05} review their properties. In galaxy groups, most

1393: lie within $\sim$300 kpc of a massive galaxy, whereas dIrr's are

1394: more dispersed \citep{Grebel05}.

1395: 

1396: SF and chemical evolution histories of Local Group dSph's can be detailed with ground-based

1397: and HST multi-colour photometry to provide CMDs, and by using large

1398: aperture telescopes to obtain intermediate-resolution spectra

1399: of $\sim$100 stars simultaneously \citep[for example][]{Venn04}.

1400: With metallicity from spectra, one can then fit isochrones to broad-band

1401: photometry to pin down the SF history. Of course, a complex SF history

1402: will superimpose multiple isochrones, increasing the likelihood

1403: of a non-unique decomposition \citep{Gallart99}, for example Figure

1404: \ref{fig:CMD-of-Carina}.

1405: 

1406: \begin{figure}

1407: \begin{centering}\includegraphics[scale=1.2]{figure7.pdf}\par\end{centering}

1408: \caption{\label{fig:CMD-of-Carina}CMD of the centre of the Carina dSph, a

1409: MWg satellite, showing multiple sub-giant branches, and blue and red

1410: HB stars \citep[][used with permission]{Monelli03}. The vertical axis

1411: is related to log visible-light luminosity, and the horizontal

1412: axis is colour with blue at left and red at right.

1413: Isochrone fits yield distinct populations

1414: aged 0.5, 5, and 11 Gyr from the present if helium and metal abundances are $Y=0.23$

1415: and $Z=0.0004$, respectively.}

1416: \end{figure}

1417: 

1418: Key results come from CMD and metallicity studies. First, except for

1419: the Ursa Minor dSph \citep{Carrerra02}, all Local Group dSph's show intermediate age stars

1420: hence prolonged SF histories \citep{Koch06,Dolphin02,Saviane00,Gallart99}.

1421: This is very surprising because energy

1422: from the first $\sim1000$ SNe in the first star burst exceeds greatly the binding

1423: energy in the shallow potential of the galaxy (total mass only $10^{7}$ \Ms),

1424: implying the ejection of all ISM then. Consider the Carina dSph,

1425: the only such system whose SF can be resolved:

1426: Figure \ref{fig:CMD-of-Carina} shows that it

1427: formed stars $<1$, $3-6$ ($\sim70\%$ of total), and $11-13$ Gyr

1428: ago with few between.

1429: Recent VLT spectra yield

1430: mean metallicity $[Fe/H]=-1.7$, ranging from -3.0 to 0. Metal-rich

1431: stars concentrate toward the centre. Carina has $\Upsilon\sim32$

1432: with uncertain DM content and essentially no gas. Multiple star bursts

1433: mean either that gas ejection was inefficient or that new gas was

1434: accreted. Accretion or re-accretion would be difficult because of this dSph's

1435: very low total mass, small cross section, and high orbital speed

1436: in the MWg's outer halo ($\sim200$ \kms). So, merging gas

1437: would have had to move in nearly the same orbit. Or, gas ejection

1438: may not remove the densest giant molecular clouds (GMC) where SF

1439: occurs. What causes the few Gyr latency between bursts is unknown.

1440: 

1441: A second indicator of a history of chemical evolution is the discovery that dSph's

1442: like Carina range substantially in chemical composition. As expected,

1443: mean metallicity increases with luminosity in dSph's. \citet{Grebel03}

1444: among others have noted that the metallicity-luminosity relation in

1445: dSph's offsets to higher metallicity at a given luminosity compared

1446: to dIrr's (gas-rich low-luminosity galaxies, generally with substantial

1447: rotation), thereby questioning if the latter can be precursors of

1448: dSph's.

1449: 

1450: Third, is the question of whether the MWg halo, particularly

1451: its component in retrograde motion, could have assembled from accreting

1452: dSph's. \citet{Venn04} have analysed detailed abundance patterns

1453: in the MWg halo and in dSph's. They find that dSph's have lower {[}$\alpha$/Fe]

1454: at given {[}Fe/H] than halo stars. There is better consistency with

1455: halo stars in extreme retrograde orbits, but other element abundance

1456: ratios do not agree. Higher {[}Ba/Y] in dSph stars compared to that

1457: in halo stars indicates slower chemical evolution in dSph's, because Ba forms in the

1458: s-process. This result accords with the extended SF histories of dSph's.

1459: So, while the halo is clearly not assembled from stars with the chemical evolution

1460: of today's dSph's, it is unclear if earlier mergers are excluded entirely,

1461: especially because the most metal-poor stars in dSph's appear

1462: to have abundance patterns more consistent with the halo.

1463: 

1464: Finally, the distribution of Local Group satellites is interesting. \citet{Hartwick02}

1465: finds evidence for dissipational collapse from the anisotropic velocity

1466: ellipsoid of this population: it is flattened spatially along the

1467: major axis, indicating a non-spherical potential. M31 and its satellite

1468: retinue including M33 show a similar alignment \citep{Koch05}. Hartwick

1469: argues that such is expected when a flattened matter filament

1470: collapses rapidly to a plane as the globular cluster and dwarf galaxy population

1471: formed. The collapse then progresses in the plane to form the giant

1472: galaxy \citep{Larson69}. \citet{Azzaro06} examine

1473: the alignment of satellite galaxies in the SDSS dataset, and find

1474: that red ones are aligned along the major axis of red hosts

1475: up to 0.5 Mpc away; the sample is too small to test for isotropy

1476: around blue hosts.

1477: 

1478: Thus there is no consensus yet on the origin of

1479: dSph's because of their surprisingly large spread in stellar age and

1480: chemical composition. The nature of these extremely

1481: low mass galaxies remains a particularly important challenge for any

1482: scenario of galaxy evolution.

1483: 

1484: 

1485: \subsubsection{\label{sub:Local Group-globular-clusters}Local Group globular clusters.}

1486: 

1487: The next rung down in mass and our main probe of baryonic halos beyond

1488: the Local Group are globular clusters. There are $\sim500$ in M31, and 

1489: $\sim150$ in the MWg \footnote{See the tabulation at

1490: \url{http://physun.physics.mcmaster.ca/Globular.html}} with continuing

1491: discoveries from near-IR surveys. Some Egs

1492: have more than 10,000. HST or an adaptive optics system resolve all but the

1493: most distant Local Group globular clusters to their core.

1494: Not all are thought to be bound in galaxies, 

1495: some may be `intergalactic tramps'. While

1496: there are counter examples of globular clusters younger than 1 Gyr in the LMC and

1497: especially in merging systems, the MWg, M31, LMC, Fornax dSph, and

1498: Sagittarius dIrr's all have globular clusters of comparable ancient age within $\pm1$

1499: Gyr. An exception is the massive cluster $\omega$ Cen, which has

1500: multiple stellar populations and anomalous chemical abundances (especially

1501: copper) compared to other globular clusters.

1502: 

1503: Globular clusters in most galaxies have a bi-modal colour distribution

1504: \citep[for example][]{Ashman92}, usually interpreted

1505: as two populations and the signature of a hierarchical

1506: build-up. However, \citet{Yoon06} have challenged this interpretation

1507: by noting the strong non-linear dependence of globular cluster colour on metallicity

1508: through the dependence of HB morphology on metallicity. They argue

1509: that the colour distribution is due instead to a continuous

1510: range of metallicity in a single old population, thereby avoiding

1511: multiple metallicity and/or age populations in globular cluster distributions.

1512: 

1513: 

1514: \subsubsection{\label{sub:Chemo-dynamical-clues-in}Local Group archaeology: chemo-dynamical

1515: clues in Local Group baryon halos.}

1516: 

1517: Local Group archaeology aims to sequence the events that formed

1518: those galaxy discs and halos, and thereby to constrain models of galaxy formation.

1519: Including recently identified objects,

1520: the MWg baryon halo has $\sim3\times10^{9}$ \Ms\  within 50 kpc

1521: radius ($<0.6\%$ of the total), density profile $\rho(r)\propto r^{-3.5}$

1522: out to radius $\sim100$ kpc, and may have two components (inner Halo

1523: I moderately flattened, outer Halo II more spherical according to

1524: \citealt{Gould98}).

1525: Analyses of both globular clusters \citep{Rodg84,Lee99}

1526: and halo field RR Lyraes \citep{Bork03,Kin07}

1527: and subdwarfs \citep{Maj96} identify two kinematic

1528: components to the halo: one in mean retrograde motion, the other in slow prograde 

1529: rotation \citep{Chiba01,Reid05}.

1530: The former is believed to result

1531: from tidal disruption of accreted satellites.  The key question of

1532: whether most of the halo formed in a single monolithic collapse,

1533: \citep{Eggen62}, or through the

1534: more stochastic accretion of fragments \citep{Searle78}, is still

1535: under debate \citep[for example][]{Chiba01,Gratton03}.

1536: 

1537: Phase space structure in the halo may be preserved, but the relative importance

1538: of dissipation over accretion during halo formation is unknown. With enough

1539: stellar kinematical data, \citet{Freeman02} argue

1540: that one can identify the stream from each accreted satellite.

1541: Few stars have been observed at sufficient spectral resolution for this, 

1542: and in any event spectra can today be obtained only

1543: of inner halo objects. In the Solar

1544: Neighbourhood only 0.5\% of the stars are in this component, a few

1545: thousand high-velocity or metal-poor stars \citep{Carney96}. Halo

1546: I and II stars are $10-13$ Gyr old, but the bulge seems to be younger,

1547: $8-11$ Gyr. There is a large range in {[}Fe/H] = -5 to +0.5, but

1548: element abundances in Halo I differ clearly from those measured in Local Group dSph's.

1549: 

1550: Photometry is much more extensive, for example from the SDSS. At $V>18$,

1551: density enhancements in Halo II are large \citep{Ibata03,Newberg02},

1552: especially tidal debris from the Magellanic Stream plus 

1553: Sagittarius dwarf galaxy

1554: \citep[our nearest neighbour galaxy,][]{Majewski03,Majewski03a,Belokurov06} 

1555: that together span the sky.

1556: A few moving groups have been identified ($1^{\circ}\times10^{\circ}$

1557: stream associated with globular cluster Pal 5, \citealt{Odenkirchen03}), and the

1558: mean and dispersions of their metallicities and kinematics have been

1559: measured; the Monoceros stream has been interpreted either as an accreted

1560: galaxy at $\sim15$ kpc Galactic radius \citep[for example][]{Helmi03,Newberg02}

1561: or as part of the MWg's warped disc \citep{Momany04}.

1562: 

1563: M31 is the other Local Group galaxy with a detectable halo (the halo of M33

1564: is elusive, \citealp{McConnachie06}, but see \citealp{Ibata07}).

1565: The $V-$ and i-band Isaac Newton Wide-Field

1566: Camera Survey of M31 resolves stars 16 times less luminous than those

1567: at the tip of the RGB and extends out to 55 kpc radii \citep{Ibata01}.

1568: The Canada-France-Hawaii Telescope MegaCam survey goes deeper and covers a larger field.

1569: In these images \citet{Ibata01} and \citet{Ibata07} find a half-dozen star streams

1570: in that bulge/halo, including one at 70 kpc (the Giant Stream, comparable to the Magellanic

1571: Stream) that extends for 100 kpc and whose large velocity deviation indicates no association

1572: with either of M31's large satellites M32 and NGC 205 \citep{Chapman06}. Assuming

1573: an orbit apocentre of 125 kpc, they find \citep{Ibata04} 

1574: $7.5_{-1.3}^{+2.5}\times10^{11}$ \Ms\  halo mass within.

1575: \citet{Guhathakurta06} study the halo at 10 \kms\ resolution and obtain metallicities

1576: from the Ca~II triplet.

1577: \citet{Chapman06} detect a distinct, non-rotating,

1578: metal poor ($[Fe/H]\sim-1.4\pm0.2$) stellar

1579: component from 10 to 70 kpc radius, of abundance similar to

1580: Halo I of the MWg \citep{Chiba01}. Evidently, M31 has had a more exciting life than the MWg,

1581: including a surprisingly recent merger (0.25 Gyr according to simulations by \citealt{Font06}).

1582: 

1583: In summary, the seeming simplicity of Egs once suggested that they

1584: are homologous in mass, largely supported by random motions but

1585: flattened by rotation, and assembled quickly

1586: (a few collapse times) while forming stars efficiently. Improved observational tools

1587: allowed study of `ellipsoids' of all 

1588: masses, and we have seen in this section that none of these assumptions have turned out to be true.

1589: 

1590: \section{\label{sec:Bulges-of-Spiral}Bulges, pseudo bulges and bars in disc galaxies}

1591: 

1592: An Sg generally contains a spheroidal stellar component that is more

1593: centrally concentrated than its flattened disc. This `bulge'

1594: provides clues for galaxy evolution. As section 3 explained, the $\Lambda$-CDM

1595: scenario assembles a bulge through a near-equal mass merger. The result

1596: is an Eg until it reacquires a disc by cooling the

1597: residual diffuse halo of ISM that was heated by the merger. Hence

1598: in the $\Lambda$-CDM scenario, Sg bulges should resemble Egs in their

1599: internal dynamics and stellar populations. So, is a massive bulge

1600: also flattened mainly by an anisotropic velocity dispersion tensor?

1601: What has the bulge of the MWg told us about other bulges?

1602: Are bulge properties consistent with hierarchical mergers, or are

1603: other formation processes relevant?

1604: 

1605: To answer, one must address two issues that regard the bulge as a

1606: separate entity from the often dusty disc. First, we must separate

1607: their light profiles. Large bulges show the centrally concentrated

1608: SB profiles that characterize Egs, a S\'ersic profile with index

1609: $n\sim4$. Disk light declines in an exponential \begin{equation}

1610: I(R)=I_{0}\exp(-R/R_{\rm{d}})\end{equation}

1611: of $I_{0}$ central SB and $R_{\rm{d}}$ scale length. Sometimes

1612: the disc dominates at large radii so can constrain $R_{\rm{d}}$

1613: while the bulge dominates at small. When separating disc and bulge

1614: light, the outer fit is then extrapolated inward to constrain $I_{0}$

1615: and is subtracted from the observed profile; the residual is fit with

1616: a S\'ersic profile to define the bulge \citep[for example][]{Schombert87,MacArthur03}.

1617: Thus, the bulge profile is only reliable for a pure exponential disc

1618: profile. 2D surface photometry is required and, in early-type (i.e.,

1619: bulge-dominant) Sgs, projection effects must be considered \citep{Noordermeer06}.

1620: Dust grains attenuate bulge starlight. In the B-band, \citet{Driver07} find that

1621: only 29\% of the bulge starlight escapes into the intergalactic medium.

1622: 

1623: \subsection{Pseudo bulges}

1624: 

1625: Second, it seems that not all bulges arise from mergers. Instead,

1626: as reviewed by \citet{Kormendy04}, `pseudo bulges' evident

1627: in disc-dominant (so-called late-type) Sgs are often built by secular 

1628: processes; \citet{Noordermeer06} finds similar flattened bulges

1629: even in some early-type Sgs. 

1630: Specifically, non-axisymmetric disturbances (bars) can trigger gas inflow and subsequent star 

1631: formation to build up a pseudo-bulge.

1632: That its scale length

1633: is well-correlated with that of the disc \citep{Courteau96,MacArthur03}

1634: further suggests its secular origin.

1635: 

1636: Other structures that at low spatial resolution mimic `classical'

1637: bulges (but resolve into more complex structures in HST and

1638: adaptive optics system images) are inner discs, bars within bars, and box-shaped nuclear

1639: bulges (perhaps nuclear bars seen edge-on). Near-IR surveys that are sensitive

1640: to the stellar mass distribution have found many bars in visible-light

1641: unbarred Sgs. For example, the bulge of M31 is barred \citep{Beaton06},

1642: and is twice as massive as that in the MWg, perhaps \citep{Brown06}

1643: having been augmented by a merger with a galaxy of MWg mass.

1644: \citet{Kormendy04}

1645: conclude that there are comparable numbers of early- and late-type

1646: galaxies, classical bulges and pseudo bulges. However, pseudo bulges

1647: in late-type Sgs have masses only $1-10\%$ of bulges in Egs.

1648: \citet{Driver07} find from their Millenium Galaxy Catalog that 60\% of stellar baryons

1649: lie in galaxy discs, and 27\% in classical bulges.

1650: 

1651: \citet{Corollo03} reviews photo-kinematical distinctions between

1652: classical and pseudo bulges, finds that they are quite blurred, and

1653: challenges their division into distinct components. To the extent

1654: that this distinction is tenable, the consensus is that 

1655: the central light excess over an inward-extrapolated exponential disc tends,

1656: in late-type Sgs, to be a pseudo bulge from secular evolution, but

1657: is a classical bulge in bulge dominant Sgs.

1658: 

1659: \subsection{Bars and other non-axisymmetric distortions}

1660: 

1661: As was touched on above, bars and other non-axisymmetric distortions

1662: can play an important r\^ole in the secular evolution of pseudo bulges.

1663: They are reviewed by \citet{Sellwood93}, \citet[focused primarily on

1664: galactic rings, but discussing bar dynamics too]{Buta96},

1665: \citet{Knapen99}, \citet{Knapen00}, \citet{Shlosman01},

1666: and \citet{Kormendy04}.  

1667: Specifically, bars drive gas inward to form stars, and they

1668: scatter disc stars into the pseudo bulge.  

1669: Here we address only

1670: questions related to the issue of bulge versus disc development:

1671: How are bars generated?

1672: How long do they last?

1673: How often are they found?

1674: 

1675: Bars form either through internal secular processes or through external

1676: tidal triggering.  Under appropriate conditions set by the mass distribution

1677: in the disc and halo, and the disc velocity dispersion, gas discs are unstable

1678: to bar and spiral structure formation through the `swing amplification'

1679: mechanism \citep[for example][]{Toomre81}.  In brief, low-amplitude leading spiral wave

1680: patterns propagate through the galaxy centre and emerge as amplified trailing 

1681: waves.  A bar develops by exchanging angular momentum with the outer stellar disc,

1682: DM halo, and/or gas disc.  In hot

1683: stellar systems (e.g., Egs and stellar interiors), any energy

1684: exchange produces a bifurcated core-halo structure \citep{Lynden68} because of

1685: the negative specific heat of a self-gravitating system.  Similarly,

1686: its angular momentum components separate \citep{Lynden72,Kormendy05},

1687: with resonance between the rotation frequency and the frequency of

1688: radial oscillation delineating the in/outward flows of angular momentum.

1689: Extensive explanations for bar formation based strictly on orbit theory are

1690: provided in, for example \citet{Sellwood93,Buta96}.

1691: Naturally, numerical simulations of spiral discs are a crucial reality check on

1692: simplified Hamiltonians.  Early simulations were restricted to 2-D, and were

1693: sufficiently grainy due to the small number of particles tracked that the secular bars

1694: that formed could plausibly have been artifacts of numerical

1695: instability.  However, with greatly increased sophisticated numerical

1696: models now in 3-D, it is clear that real bar instabilities

1697: occur and that their strength depends on the mass distribution and velocity dispersion of the disc

1698: and halo \citep[for example][]{Knapen99,Shlosman01}.

1699: 

1700: While bar instabilities may form in isolated discs only through internal processes,

1701: tidal interactions and minor mergers are also effective triggers to

1702: drain angular momentum from the gas and inflow it

1703: to perhaps form stars and build up a pseudo bulge.  Thus

1704: it is of considerable interest whether most pseudo bulges have been created by

1705: external (interaction/merger) processes, which would be more

1706: consistent with the hierarchical merger picture \citep{Kannappan04},

1707: or by internal processes, which would pose problems for that

1708: scenario \citep{Kormendy05}.  In support of the external trigger, \citet{Kannappan04}

1709: find a correlation between central blue colours in pseudo bulges

1710: with morphological evidence of tidal encounters or merging.

1711: On the other hand, \citet{Kormendy06} examined two

1712: pseudo bulges in detail, and found no evidence for an external trigger.  

1713: 

1714: The longevity of a bar depends upon the importance of

1715: gas in its evolution \citep[for example, the simulations of][]{Debatt06}.  

1716: For purely stellar systems, bars

1717: tend to be long-lived, in fact are only weakened by tides \citep{Athanassoula03b}.

1718: This is because angular momentum flows

1719: from the bar to either the outer disc or the DM halo, which slows the

1720: pattern speed and increases the strength of the bar

1721: \citep[for example][]{Athanassoula03a,Athanassoula03b}.

1722: On the other hand, when gas is included, there

1723: is general consensus that its interaction with the bar weakens

1724: the bar, but little agreement on how this occurs.

1725: \citet{Bournaud05} propose that the bar drains

1726: angular momentum from the gas, which increases bar pattern speed and weakens the

1727: bar.  \citet{Berentzen07} argue that the gas has little direct effect.

1728: Instead, the bar is destroyed by the build up of the pseudo bulge

1729: as gas inflows to form stars, a less direct r\^ole for gas than 

1730: \citet{Bournaud05} propose.  The

1731: timescale for bar destruction is similarly unclear, with some investigators

1732: finding a quick $\sim$1-2 Gyr one \citep{Friedli94,Bournaud05},

1733: and others much longer \citep[for example][]{Berentzen07}.  

1734: 

1735: Finally, observations in the near-IR H band, which better reveal the bulk of

1736: the underlying stellar distribution than do optical passbands, find that

1737: $\sim2/3$ of Sgs show a bar

1738: \citep{Knapen99,Knapen00,Eskridge00,Whyte02,Marinova07}.  

1739: In addition, bars are more common among galaxies with a central starburst 

1740: \citep[for example][]{Hunt99}, but whether

1741: they are more common in galaxies with AGN than those without remains controversial 

1742: \citep[for example][]{Mulchaey97,Knapen00,Laurikainen04}.

1743: 

1744: \subsection{Constraints from stellar populations}

1745: 

1746: Stellar population studies consider mostly visible-light

1747: and near-IR colours, not line strengths. As noted in section \ref{sub:Results-for-the},

1748: combined visible-band and near-IR colours break the age-metallicity

1749: degeneracy. However, dust in Sgs dust extinguishes starlight to introduce

1750: a new degeneracy. Near-IR colours are relatively immune to dust, but

1751: their combination with dust-sensitive visible-band colours

1752: ameliorates the age-metallicity degeneracy. To avoid uncertain

1753: reddening and contamination by young stars in a superimposed disc,

1754: \citet{Peletier99} studied nearly edge-on Sgs near

1755: $\sim R_{\rm{e}}$. They found that

1756: the massive bulges of early-type Sgs resemble those of early-type

1757: galaxies in the Coma cluster. Specifically, derived ages at $\sim R_{\rm{e}}$

1758: scatter by $\lesssim2$ Gyr around $\sim$10 Gyr mean; the absolute

1759: age is uncertain from uncertain modelling of near-IR colours. In contrast,

1760: smaller bulges in late-type Sgs are bluer, hence younger. As for

1761: Egs, where more massive galaxies are older in the mean, the trend

1762: of older more massive bulges seems to contradict the prediction that

1763: hierarchical mergers form smaller things first. However, one must

1764: recall that smaller Sg bulges are probably dominated by secular

1765: processes.

1766: 

1767: When examined spectroscopically, the centres of Sg bulges are found

1768: by \citet{Proctor02} to be younger and sparser in $\alpha$-elements

1769: than Egs. \citet{Moorthey05} find that large bulges in Sgs resemble

1770: massive Egs, both being old and red, with enhanced $\alpha$-elements.

1771: \citet{Norris06} obtained long-slit spectra to $\sim2R_{\rm{e}}$

1772: in the edge-on S0 galaxy NGC 3115 and find an old, light-weighted

1773: mean age for its bulge, while the disc is $\sim5-8$ Gyr old.

1774: Given the confused disentanglement of bulge and disc light \citep{Corollo03},

1775: detailed maps of spectral line indices over a wide range of `bulge'/disc

1776: contributions are necessary to interpret indices obtained

1777: from sparse slits on each galaxy.

1778: 

1779: Dust restricts our view of the MWg's inner bulge in visible-band light

1780: to `Baade's Windows', a few sightlines of reduced extinction.

1781: In the near-IR, extinction is hugely reduced to reveal a clear triaxial

1782: bulge in for example the \textit{COBE} image \citep{Dwek95}. \citet{Babusiaux05}

1783: summarize recent work that shows the triaxial bar ending in a possible

1784: inner ring whose longest axis is inclined to our l.o.s.\ by $\sim22^{\circ}$.

1785: This barred inner bulge, prominent in counts of RC stars, appears

1786: to be secular in origin. So, it is surprising and perhaps problematic

1787: for the secular evolution picture that stars in Baade's largest `Window'

1788: (at Galactic latitude $-4^{\circ}$) are old \citep{Zoccali03,Kuijken02}

1789: and have enhanced $\alpha$-element/Fe abundances \citep{McWilliam94,Barbuy99}.

1790: near-IR imaging spectroscopy of this region will continue to be important

1791: for understanding secular bulges in general and the stellar populations

1792: of the MWg bulge in particular.

1793: 

1794: 

1795: 

1796: 

1797: \section{\label{sec:Thick-&-Thin}Disk Systems }

1798: 

1799: 

1800: \subsection{Luminous Structures}

1801: 

1802: Aside from the aforementioned (pseudo-)bulge population, and a small

1803: halo population (as evidenced by metal-poor stars and globular clusters in the MWg

1804: and by globular clusters in other galaxies), Sgs are characterized by a prominent

1805: disc. Disks are highly dissipated, kinematically cold (rotating) structures,

1806: conventionally assumed to have forgotten their progenitors, despite having made only a

1807: few tens of rotations since their formation. 

1808: 

1809: The azimuthally

1810: averaged SB profiles of discs are exponential in both radial and vertical

1811: directions.

1812: Many discs are dusty;

1813: \citet{Driver07} find that 63\% of B-band photons released in discs actually escape from

1814: these galaxies.

1815: As was discussed in section 4,

1816: discs suffer non-axisymmetric instabilities that can develop into distortions, 

1817: predominantly two-armed spirals. The nature of spiral structure

1818: is a major subject discussed by, for example \citet{Bertin96}. 

1819: Here we concentrate just on issues that relate directly to

1820: the formation and evolution of whole galaxies.

1821: Specifically, we address

1822: 

1823: \begin{itemize}

1824: \item The existence of thin and thick discs, and the relationship between

1825: these apparently distinct components.

1826: \item That both ionized and neutral gas can trace Sg discs out to large

1827: radii, enabling a more definitive measure of the DM in Sgs than is

1828: possible in Egs.

1829: \item That Sg discs harbour most of the ongoing SF in the universe, making

1830: them important laboratories for studying SF and feedback processes.

1831: \end{itemize}

1832: 

1833: \subsubsection{Thick disks.}

1834: 

1835: \noindent \citet{Burstein79} and \citet{Tsikoudi79} discovered that

1836: the vertical distribution of light in some edge-on galaxies has two

1837: exponential components; that with larger scale height is the thick

1838: disc. Later, \citet{Gilmore83} established a thick disc in the MWg

1839: by counting stars by apparent brightness perpendicular to the Galactic

1840: plane. Such counts are the convolution of the vertical density function

1841: with the very broad stellar luminosity function, so are hard to interpret

1842: uniquely because any feature in the luminosity function can be misinterpreted

1843: as a feature in the density function. The stellar luminosity function

1844: is broad both because of the mass-luminosity relation for MS stars

1845: and because evolved stars range widely in luminosity. The thin disc

1846: has $280-350$ pc scale height at $R_{0}$ and contains most of the

1847: baryonic angular momentum. The thick disc has $0.9-1.2$ kpc scale

1848: height and $\rho_{\rm{thick}}/\rho_{\rm{thin}}=0.085_{-0.065}^{+0.045}$

1849: locally \citep{Siegel02}. The two discs overlap maximally at $1-1.5$

1850: kpc at $R_{0}$; sparse, faint stars complicate their separation and

1851: normalization. Adding colours improves deconvolution prospects.

1852: 

1853: It is now customary to separate the discs kinematically \citep[for example][]{Carney96},

1854: a statistical separation because assigning any star to one disc is

1855: uncertain. While, on the whole, stellar metallicity declines from

1856: the thin disc to thick to halo (the thin disc has {[}Fe/H] = -0.7

1857: to +0.3 and the thick has {[}Fe/H] = -2.2 to -0.1), there is enough

1858: overlap that again only statistical assignment is possible. \citet{Reddy06}

1859: have observed a sample of $\sim100$ thick disc stars, plus fewer

1860: stars in the thin disc and halo. It is clear that thick and thin disc

1861: stars have strikingly different abundance patterns with little overlap.

1862: Specifically, $\alpha$-elements are relatively enhanced in thick

1863: disc stars. In addition, these stars are on average 5 Gyr older.

1864: 

1865: Suggested origins of the thick disc include a population of dissolved

1866: super-star clusters of $\sim10^{6}$ \Ms\ \citep{Kroupa02},

1867: or the infall of satellites \citep{Bekki00}. \citet{Reid05} argues

1868: from the white dwarf luminosity function for a continuous sequence of formation

1869: from halo ($1\%$ of white dwarfs in the Solar Neighbourhood) to thick ($20\%$)

1870: and ultimately thin discs ($\sim80\%$). The low-luminosity cutoff

1871: of the white dwarf cooling curve tells us that the thick disc is $\sim9.5\pm1$

1872: Gyr old, with most of its stars having formed over $\sim1$ Gyr; it

1873: has {[}Mg/Fe] $>$ 0.2 (average in one sample {[}Fe/H] = -0.49). Its

1874: abundances overlap with those of metal-poor stars in the thin disc. Cooling

1875: curves also show that the thin disc began to form many stars $7-8$

1876: Gyr ago. An age-metallicity relation is evident in the thick disc:

1877: {[}Fe/H] increases by $\sim0.5$ as the age decreases by $\sim5$

1878: Gyr. \citet{Reddy06} note that chemical abundance patterns in thick

1879: disc stars are distinct from those of stars in dSph's and in the Magellanic

1880: Clouds \citep{Shetrone03,Venn04}. Thus, the thick disc cannot have

1881: been built from mergers of satellites that resemble our surviving

1882: current dSph population. Moreover, the distinct abundance patterns

1883: of thick and thin discs exclude the possibility that the thick disc

1884: could be a dynamically heated thin disc \citep[for example][]{Feltzing04}.

1885: 

1886: \citet{Bland-Hawthorn04} propose new instruments on 8-meter aperture

1887: telescopes to `chemically tag' stars, to learn the dissipation

1888: history of the MWg's disc. Many elements cannot form in normal stellar

1889: evolution, so reflect the chemistry of the progenitor ISM.  Stars

1890: of expected similar age in the Solar Neighbourhood travel in distinct

1891: kinematical `moving groups', suggesting diffusion from common

1892: birth sites. Tagging aims to use detailed abundance patterns that

1893: are correlated with kinematical streams to track stars with identical

1894: abundances back to specific formation sites. In the first test of

1895: homogeneity, \citet{deSilva05} find negligible intrinsic scatter

1896: in heavy element abundances (Zr, Ba, La, Ce, Nd) among Hyades cluster

1897: members. The key uncertainty to resolve by study of other clusters

1898: is if there is sufficient

1899: chemical diversity across the disc to tag all sites uniquely.

1900: 

1901: Thick discs are ubiquitous.

1902: While some early studies found them in early-type Sgs but not in types

1903: Sc and later, \citet{Yoachim06} have found them in virtually all Sgs sampled.

1904: They argue that thick discs result from early mergers of satellites. 

1905: 

1906: 

1907: \subsubsection{Warps and the outer edges of discs.}

1908: 

1909: Outer discs are seldom flat. The warp of the MWg was well described by 

1910: \citet{Oort61} and extragalactic warps were mapped in HI with early radio interferometers

1911: \citep[for example, M33, by][]{Rogstad76}.

1912: As sensitivity improved, photometry of edge-on \citep[for example][]{Garcia02}

1913: Sgs and the rotation curves of highly inclined Sgs \citep[for example][]{Noordermeer07}

1914: often show warps in the HI disc beyond the starlight. For example,

1915: the HI layer of the MWg thickens monotonically from $R_{0}$, and

1916: beyond $1.5R_{0}$ warps up to $\sim6$ kpc ($\pm25^{\circ}$) from

1917: the disc plane on one side at 30 kpc radius (the other side is much

1918: flatter) \citep{Levine06}. Stars follow the warp.  Adding a tidal

1919: wake in DM to the direct tides from the Magellanic Clouds \citep{Hunter69},

1920: \citet{Weinberg06} explain most of these aspects by tides that

1921: resonate in the outer gas disc of the MWg; the $m=2$ mode is an important

1922: constraint that bears on dynamical alternatives to DM and on the shape

1923: of the DM halo. The warp is very dynamic, `like a flag flapping

1924: in the breeze'. The two other Local Group Sgs have prominent, asymmetric

1925: warps that may arise from mutual tidal interaction, triaxial DM halos,

1926: or accretion of cold gas.

1927: 

1928: While some discs truncate at $\sim4$ exponential

1929: scale lengths \citep{deGrijs01}, others do not \citep[for example][]{Erwin05}.

1930: Deep images of NGC 300 \citep{BH05}

1931: show stars out to 10 radial scale lengths. The discs of a few

1932: low-SB galaxies (section \ref{sub:Low-surface-brightness}) are seen

1933: beyond 30 scale lengths, including the prototype of this class, Malin 1, which was

1934: found by \citet{Barth07} to have a normal disc component

1935: embedded within the 100 kpc extended component.

1936: At such depths, modelling the contamination

1937: of starlight by the noise of compact background galaxies is critical;

1938: assumptions have been validated against deep galaxy counts

1939: from HST.

1940: 

1941: \subsection{Global scaling relations: Mass-Velocity correlation in Sgs}

1942: 

1943: Section 3 showed that constraining Eg internal dynamics is frustratingly

1944: difficult. The tensor Virial Theorem shows that massive Egs are flattened by velocity

1945: anisotropy not rotation, but we have much to learn about their internals.

1946: Considering the homology of global properties

1947: of Egs leads to the Fundamental Plane, a global spatio-kinematical

1948: correlation. 

1949: 

1950: However, Sgs provide a more favourable situation

1951: for deciphering internal dynamics and mass distributions because

1952: kinematical tracers in the cold, rotating disc can be measured out to

1953: many radial scale lengths. We first consider global scaling relations

1954: for Sgs, in analogy with study of Egs that uncovered the FP. In

1955: section \ref{sub:Dark-matter-content1}, we consider what can be learned

1956: about the distributions of light and dark matter from fits to

1957: the measured radially variable rotation speed of the disc (the

1958: rotation curve, RC). We consider what can be inferred on internal

1959: properties from an optimum RC for a single galaxy as opposed to statistical

1960: analysis of a sample.

1961: 

1962: Sg RC's are measured with emission lines from regions of

1963: ionized hydrogen, and with the $\lambda21$ cm hyperfine spin-flip

1964: transition in the ground-state of hydrogen (HI). The latter line

1965: is unobscured by dust, and \citet{Tully76} measured its full width integrated

1966: across an Sg disc to establish a correlation between the rotation

1967: speeds of discs and their B-band luminosities. This Tully-Fisher relation

1968: (TFR) is the Sg analogy and precedent of the FP. Its small scatter

1969: spurred work on the extragalactic distance scale, galaxy peculiar

1970: velocities, and 3D cartography of the local universe. Here, we consider

1971: what can be inferred about homology of the global properties of Sgs

1972: from the existence of, and small scatter in, the TFR.

1973: 

1974: The TFR establishes that the disc rotation speed (section \ref{sub:Evidence-from-other})

1975: --- defined optimally \citep{Verheijen01,Noordermeer06} by the `flat

1976: part' of the RC (although RC's are not particularly flat, section 5)

1977: --- correlates with total baryonic mass --- as measured

1978: originally from the B-band luminosity and now including ISM \citep{McGaugh01}

1979: to account for all baryons.

1980: The tightest correlation

1981: uses K-band luminosity and HI to map baryonic mass while minimising

1982: corrections for dust extinction and variable SF to find \begin{equation}

1983: V_{\rm{flat}}\propto L_{\rm{K}^{\prime}}^{x}(\rm{baryonic}),\end{equation}

1984: with \citep{Verheijen01,Noordermeer06} $x=4$. One converts to mass

1985: with an assumed $\Upsilon_{\rm{d}}\equiv[M/L]_{\rm{d}}$. $\Upsilon_{\rm{d}}$

1986: is constrained by galaxy disc colours (from, for example \citealt{Bell01}'s

1987: grid of stellar population models, to yield $x\sim3.5$) and by a

1988: universal mass-discrepancy relation \citep{McGaugh04} that is consistent

1989: with both colours and standard IMF's. $\Upsilon_{\rm{d}}$ seems

1990: to vary little among different Sgs \citep{Bell01,McGaugh04}.

1991: 

1992: This tight correlation, coupled with its independence of parameters

1993: such as disc size and mean SB \citep[for example][]{Courteau99},

1994: shows that the internal dynamics of Sgs do not depend on how baryons

1995: are distributed within the potential. Indeed, \citet{McGaugh05a} finds

1996: the same behaviour in low-SB galaxies (LSBg's) --- those whose average

1997: SB is fainter than 40\% of the moonless night sky \citep{Bothun97,Impey97}.

1998: We adopt the conventional explanation that the fundamental homologies

1999: of Sgs arise from ubiquitous DM halos, with baryons simply `garnishing'

2000: the result. An alternative interpretation, MOdified Newtonian Dynamics

2001: \citep[reviewed by][]{Sanders02}, proposes to alter Newtonian acceleration

2002: at the small value reached toward the visible-band edges of galaxy

2003: discs where RC's tend to flatten. Whether or not MOND is real physics, \citet{McGaugh04}

2004: finds that its dependence on acceleration gives the tightest possible

2005: description of the discrepancy between baryons and Sg RC's, with scatter arising

2006: only from observational uncertainties.

2007: 

2008: \begin{figure}

2009: \begin{centering}\includegraphics[scale=2]{figure8.pdf}\par\end{centering}

2010: \caption{\label{fig:Space-density-of}Number density of galaxies versus

2011: their central SB $\mu_{0}$. LSBg's start

2012: slightly to the right of the vertical grey line (the SB of the darkest

2013: night sky), and extend to the left where their outer regions can currently be detected down

2014: to $\sim10^{-3}$ of this limit. They can have the same blue luminosity

2015: as high SB galaxies, but span a much larger area on the sky. Freeman's selection

2016: for high SB galaxies is at right.

2017: This figure

2018: originally appeared in the Publications of the Astronomical Society of the

2019: Pacific \citep[PASP, 109, 745]{Bothun97}.  Copyright 1997,

2020: Astronomical Society of the Pacific; reproduced with permission of the

2021: Editors.}

2022: \end{figure}

2023: 

2024: 

2025: LSBg's are the strongest test of the homology of Sgs through

2026: the TFR; they make up to 30\% of the gaseous mass of the universe,

2027: a negligible fraction of its stellar mass \citep{Driver07},

2028: and at least half of the total galaxy population (Figure \ref{fig:Space-density-of}).

2029: While mostly detected as gas-rich Sgs in the field (a few being

2030: AGN hosts) and in poor, spiral-rich clusters, a few dwarf Egs of low SB

2031: have been found. LSBg's challenge $\Lambda$-CDM by testing the proposition

2032: that baryonic luminosity functions are biased toward high-SB galaxies.

2033: To focus on this problem, we now consider Sg RC's and DM in

2034: more detail.

2035: 

2036: 

2037: 

2038: \subsection{\label{sub:Dark-matter-content1}Dark matter content}

2039: 

2040: 

2041: \subsubsection{\label{sub:Evidence-from-other}Evidence from galaxy rotation curves.}

2042: 

2043: Contrasting the limited work on DM in Egs, the RC's of gas in the

2044: discs of spiral and S0 galaxies provide compelling and direct evidence of

2045: DM \citep[for example][]{Broeils92,Cote00} because, at sufficient

2046: radius, the DM halo mass is responsible for most of the observed rotation.

2047: While evidence for important DM originated from RC's of ionized gas

2048: in regions of active SF \citep{Rubin78} (HII regions in astronomical

2049: jargon), ambiguity lingers \citep[for example][]{Kent86} because

2050: this gas is generally too faint to see at radii large enough to pin

2051: down the DM halo. Specifically, one cannot distinguish between a maximal-mass

2052: stellar disc (i.e., the largest $\Upsilon_{\rm{d}}$ whose rotational

2053: contribution does not exceed the full RC; \citealp{Sackett97,Palunas00,Weiner01,vanAlbada86})

2054: and a sub-maximal disc of lower $\Upsilon_{\rm{d}}$ ($\sim6.3$

2055: according to \citealt{Bottema93}'s study of stellar velocity dispersions)

2056: plus a more massive DM halo \citep[for example][]{Courteau99,Pizagno05}.

2057: This stellar disc --- DM halo mass degeneracy exists from S0 to Sd

2058: galaxy types.

2059: 

2060: While the radial SB profile tracks disc baryons,

2061: normalization by $\Upsilon_{\rm{d}}$ depends on vagaries of

2062: stellar population models and, in discs, on the inevitable presence of dust.

2063: Light comes mainly from the upper IMF,

2064: in discs from young hot stars that lie near dusty regions. But mass comes from numerous low mass

2065: stars. Uncertainty in the universality of an IMF established from

2066: the Solar Neighbourhood makes $\Upsilon_{\rm{d}}$ hence disc mass

2067: uncertain, thence the rest of the RC that is ascribed to the DM halo.

2068: 

2069: Evidence for DM and its isolation from baryons is compelling in galaxies

2070: with the most extended HI, where most maps from radio interferometers

2071: such as the \textit{VLA} and Westerbork Synthesis Radio Telescope show

2072: a flat or increasing RC at the largest measurable radius

2073: \citep{Rogstad78,Bosma81,Noordermeer07}.

2074: If the RC spans sufficient radii, one can see variations in the

2075: DM/baryon ratio.

2076: To establish their individual contributions

2077: one can then assume constant $\Upsilon_{\rm{d}}$ (questionable)

2078: for baryons, and a simple radial distribution of DM, usually an

2079: isothermal sphere or the cuspy \citet[NFW]{Navarro97} halo profile 

2080: motivated by simulations (section 5.3.1) but with little observational

2081: support.

2082: In analogy to the relative clarity of bulge-disc decompositions when

2083: the observed SB profiles span

2084: both bulge- (inner) and disc-dominated (outer) regions, RC's that

2085: span radii from where bulge+disc dominate (inner) to where DM halo

2086: dominates (outer) allow robust decomposition of baryonic (bulge+disc+gas)

2087: and DM components. 

2088: 

2089: RC's from long-slit spectra often jump to 200-300 \kms\

2090: within a few hundred pc radii.

2091: The rise to and location

2092: of peak rotation correlates with starlight concentration at the

2093: nucleus, arguing that bulge mass dominates dynamics \citep{Noordermeer07}.

2094: RC's often decline 10-20\% from the peak before flattening,

2095: falling fastest in luminous galaxies. For example, Figure \ref{fig:The-rotation-curve}a

2096: shows \citet{Carignan06}'s data and maximal-disc decomposition of M31 within

2097: 35 kpc radius, which yields $M_{\rm{stellar}}=2.3\times10^{11}$ \Ms, 

2098: $M_{\rm{HI}}=5.0\times10^{9}$ \Ms, and $M_{\rm{dm}}=1.1\times10^{11}$ \Ms; 

2099: this galaxy has negligible HI inside 5 kpc radius. DM dominates

2100: only at $\gtrsim30$ kpc. 

2101: 

2102: \begin{figure}

2103: \begin{centering}\includegraphics[scale=0.41]{figure9.pdf}\par\end{centering}

2104: \caption{\label{fig:The-rotation-curve}RC decompositions of (a)

2105: M31 (grey line and \fullcircle\ \citealt{Carignan06}) with a central 

2106: $10^{8}$ \Ms\ supermassive black-hole (\longbroken) \citep{Bender05} and bulge stars (\full); 

2107: (b) a high SB galaxy (NGC 2403, black lines and \fullcircle) and a LSBg

2108: (UGC 128, grey lines and \fullcircle, adapted from

2109: \citealt{deBlok96}). Shown in both panels are DM halo (\chain), stars in the

2110: disc (\broken) and gas (\dotted).}

2111: \end{figure}

2112: 

2113: Indeed, the trend is for high luminosity discs like that of M31 to

2114: be maximal in stars, whereas low luminosity galaxies and LSBg's are

2115: dominated by DM \citep[for example][]{deBlok97,Kranz03}. 

2116: Figure \ref{fig:The-rotation-curve}b compares

2117: mass decompositions of high- and low-SB galaxies.

2118: The DM halos (\chain) are very similar despite completely different baryon

2119: curves, and DM dominates throughout the LSBg. \citet{Oneil04}

2120: map in HI nearly half of 81 LSBg's that have HI mass $>10^{10}$ \Ms; 

2121: all RC's rise out to the last point measured. LSBg's have the

2122: largest $\Upsilon_{d}$ measured. Their RC's \citep{McGaugh01} are

2123: shallower, baryon surface densities are smaller at small radii

2124: than those of high SB galaxies, and are best fit with minimal discs.

2125: 

2126: Flat RC's imply that the DM density must decline no faster than $1/r^{2}$.

2127: Although no erstwhile `flat' curve has yet been seen to decline

2128: at largest radius, the DM profile must eventually steepen to yield

2129: finite mass. Halo extents probed with weak gravitational lensing \citep{Hoekstra04}

2130: and faint H$\alpha$ emission \citep{Bland-Hawthorn97} both suggest

2131: steep DM profiles beyond radii traced by HI. \citet{Persic96}

2132: argued that all (mostly visible-light from ionized gas) RC's measured

2133: by \citet{Mathewson92} outside the bulges of Sb-Sd galaxies

2134: can be explained by a Universal Rotation Curve (URC) set only by galaxy

2135: luminosity. Once a maximal stellar bulge contribution is removed,

2136: \citet{Noordermeer07} find that agreement with the URC is often,

2137: but not always, very good. At larger radii, all curves flatten but

2138: are too complex to be described by a URC that depends only on galaxy

2139: luminosity.

2140: 

2141: A recent estimate of mass distributions in Sg discs and DM halos

2142: from the best RC's is \citet{Noordermeer06}, who analysed HII and

2143: HI (out to 20 scale lengths) curves and multicolour surface photometry of 19 S0-Sab

2144: galaxies. 

2145: Rotational velocity totals 

2146: \begin{equation}

2147: V_{\rm{circ}} = 

2148: \sqrt{\Upsilon_{\rm{b}}V(r)+\Upsilon_{\rm{d}}V_{\rm{d}}^2(r)+\eta V_{\rm{HI}}^2(r)+V_{\rm{DM}}^2(r)+V_{\rm{pm}}^2(r)}

2149: \ \ \label{eq:rotation}\end{equation}

2150: \begin{figure}

2151: \begin{centering}\includegraphics[scale=5.1]{figure10.pdf}\par\end{centering}

2152: \caption{\label{fig:rotcurvedecomp}\citet[][used with permission]{Noordermeer06}'s 

2153: mass decompositions

2154: of SA galaxy IC 356, using an isothermal DM halo. (Similar results

2155: are obtained using an NFW DM profile.) The top two

2156: rows assume maximal and minimal mass stellar discs, respectively; the DM

2157: contribution is small for the former and large for the latter.

2158: the third assumes constant $\Upsilon$, and the bottom assumes a scaled

2159: HI disc of (baryonic) DM and \textbf{no} CDM halo. The left column

2160: shows the central region with \opensquare showing kinematical

2161: long-slit measurements from the ionized gas, the right shows the full

2162: RC from HI spectra; the curve width in grey shows the range spanned

2163: by the azimuthal average. Bumps and wiggles arise from non-circular

2164: motions associated with spiral arms and their resonances. Decompositions are plotted with

2165: the following lines: stellar bulge (\full) and

2166: disc (\longbroken), gas disc (\dotted),

2167: central supermassive black-hole (\broken), and DM halo (\chain).

2168: Bold (\full) is the combined model RC (\ref{eq:rotation}).}

2169: \end{figure}

2170: with $\eta=1.4$ accounting for the mass of hydrogen and other abundant

2171: elements (gas accounts for only 4\% of the total mass), $V_{\rm{pm}}(r)$

2172: the contribution of a central point mass (perhaps a supermassive black-hole), and the

2173: bulge $(M/L)_{\rm{b}}\equiv\Upsilon_{\rm{b}}.$ Both $\Upsilon_{\rm{d}}$

2174: and $\Upsilon_{\rm{b}}$ are held constant, despite bluer discs

2175: at larger radii that may result from different age/metallicity.

2176: \citeauthor{Noordermeer07} derived mass models for 17 galaxies; Figure

2177: \ref{fig:rotcurvedecomp} shows decompositions for

2178: IC~356. A maximal stellar bulge contributes little farther out; the

2179: main fitting parameters are the disc scale length and an uncertain

2180: $\Upsilon_{\rm{d}}$ that yields an uncertain DM halo profile. Nevertheless,

2181: the outermost 20 kpc of this RC (and others in his dataset) show clear

2182: need for a DM halo. The bottom panel is the best fit without a halo,

2183: boosting disc mass density $10-50$ fold \citep[using an additional baryonic disc component,][]{Pfenninger05}.

2184: While residuals are within uncertainties and the fit at $\lesssim30$

2185: kpc reproduces RC bumps and wiggles of unknown dynamical significance,

2186: it is clearly untenable beyond. (He notes that ionization by the metagalactic

2187: radiation field may decouple the DM from gas, \citealt{Bland-Hawthorn97}.)

2188: 

2189: A satisfactory fit uses a DM halo with isothermal profile and a maximal

2190: stellar disc (top panel) of larger $\Upsilon_{\rm{d}}$

2191: than expected from visible-band colours. An alternative, excellent

2192: fit (second panel from top) uses a \textit{minimal} stellar disc with

2193: reasonable $\Upsilon_{\rm{d}}$ and an NFW DM profile. Hence,

2194: definitive assessment of the DM halo profile, and its contribution

2195: relative to the disc at maximum rotation speed, is possible only if

2196: $\Upsilon_{\rm{d}}$ can be found independently. As Noordermeer

2197: notes, the most direct constraint on disc mass would come from the

2198: velocity dispersion of stars in the outer disc. One would need the

2199: very largest telescopes to measure the $\sim10$ \kms\

2200: velocity dispersion of stars at that very low SB. His minimal

2201: discs are consistent with simulated DM halos. But, his data

2202: show a larger range of central concentrations and his fitted DM halo

2203: is therefore more concentrated than simulations find. He finds that core radius

2204: increases with scale length, independent of the choice

2205: of minimal/maximal discs. 

2206: 

2207: \citeauthor{Noordermeer06}'s study shows the advantages of pursing

2208: RC's to the largest possible radius. A complementary approach is to

2209: constrain DM and baryonic contributions through statistical analysis

2210: of many Sg RC's that range over galaxy mass, SB, and visible-band

2211: diameter. Differing contributions of disc or DM halo to the maximum

2212: rotation velocity give different patterns of correlated residuals

2213: around the mean TFR. For example, \citet{Courteau99} studied a sample

2214: of luminous, non-barred galaxies. They found little correlation between

2215: residuals from their version of the TFR, specifically $\Delta\log V_{2.2}$

2216: versus $\Delta\log R_{\rm{exp}}$, with $R_{\rm{exp}}$ the scale

2217: length of the exponential disc and $V_{2.2}$ the velocity at $2.15R_{\rm{exp}}$

2218: where rotation peaks in a pure exponential disc. If discs are maximal

2219: with no DM halo inside $2.15R_{\rm{exp}}$, residuals should correlate

2220: with $\Delta\log V_{2.2}/\Delta\log R_{\rm{e}}=-0.5$. From this

2221: discrepancy and the behaviour of other residuals, they conclude that

2222: the DM halo contributes 60\% of the total mass within $2.15R_{\rm{exp}}$,

2223: so even high SB galaxies are not maximal discs at that radius. This largely

2224: model independent result has been further reinforced in barred Sgs

2225: \citep{Courteau03}. 

2226: 

2227: 

2228: \subsubsection{\label{sub:Evidence-in-the}Evidence in the MWg.}

2229: 

2230: We are supposedly surrounded by DM, but tracers through the halo are

2231: hard to identify. \citet{Ollig01} note another frustration:

2232: the local RC of the MWg $\Omega(R_{0})$ is uncertain. Errors on

2233: `consensus' values are overwhelmingly systematic, so their propagation

2234: is complicated. Thankfully, parallaxes of stars orbiting close to

2235: the supermassive black-hole at the centre of the MWg will soon measure $R_{0}$ to dispense

2236: with some of these irritations.

2237: 

2238: 

2239: \paragraph{Microlensing surveys}

2240: 

2241: Light variability curves of background stars being microlensed by

2242: MWg halo objects have been obtained by the MACHO \citep{Alcock00}, OGLE

2243: \citep{Udalski94}, and EROS \citep{Aubourg93} projects

2244: as they stare at the LMC (MACHO) or at the MWg bulge (others). Microlensing

2245: measures directly the mass distribution along the l.o.s. As the Earth's

2246: orbital motion causes a parallax shift in the observed light curve, the duration

2247: of the microlensing event (the time to cross the Einstein ring of

2248: angular width $\propto\sqrt{m}$) is set by three parameters that

2249: are degenerate observationally: the transverse velocity of the lens

2250: to us, the ratio of lens to source distances, and the lens mass $m$.

2251: 

2252: Surveys of microlensing toward the LMC have produced somewhat inconsistent

2253: results and contradictory interpretations. MACHO toward the LMC finds

2254: fewer events than expected if the total MWg halo mass arose solely from

2255: compact objects of mass $m$. Lensing optical depths $\tau\sim3\times10^{-7}$

2256: based on $\sim13-17$ events suggest that 20\% of the MWg halo mass

2257: is in $0.5_{-0.3}^{+0.2}$ \Ms\  objects \citep{Alcock00,Bennett05}.

2258: However, \citet{Novati06} proposes that many MACHO events are actually

2259: self-lensing by the LMC halo, thereby lowering the estimated DM halo

2260: fraction. Moreover, the EROS-2 collaboration finds a microlensing

2261: $\tau$ only 10\% that of MACHO \citep{Jetzer05}. Proposed new microlensing

2262: surveys will eventually resolve these controversies.

2263: 

2264: The OGLE/EROS dataset of more than 100 events provides the average

2265: $\tau$ along several sightlines toward the MWg bulge and the distribution

2266: of event durations. Several groups \citep[for example][]{Bissantz02}

2267: have used these data with far-IR $\lambda240\,\mu$m maps of the MWg

2268: bulge along sightlines through the stellar bar to argue for a near

2269: maximal stellar disc, as found in other Sgs (see section \ref{sub:Evidence-from-other}).

2270: For example, \citet{Bissantz02} reproduce the MWg RC out to 5 kpc

2271: radius without a DM halo, unsurprising given the small radius probed,

2272: and use a near-maximal stellar $\Upsilon_{\rm{d}}$. Streaming motions

2273: associated with a triaxial bar account naturally for the $>40\%$

2274: fraction of lensing events that exceed 50 days \citep{Evans02}. In

2275: short, while microlensing surveys of the bulge have not located a

2276: DM halo, they have helped to constrain the inner structure of the

2277: MWg.

2278: 

2279: 

2280: \paragraph{Stellar densities: disc and halo populations}

2281: 

2282: The local mass density can be estimated from the Jeans equation by

2283: assuming that the disc potential is steady state, and is separable

2284: into $r$ and $z$ motions because vertical variations in mass exceed

2285: greatly radial ones. One then obtains\begin{equation}

2286: \frac{\partial}{\partial z}\left[\frac{1}{\nu}\frac{\partial(\nu\overline{v_{\rm{z}}^{2}})}{\partial z}\right]=-4\pi G\rho(R_{0},z)\end{equation}

2287: with $R_{0}=8.0\pm0.4$ kpc the solar radius. Measuring the \textit{volume}

2288: variations of star density $\nu(z)$ and $\overline{v_{z}^{2}}(z)$

2289: above the MWg plane, yields $\rho(R_{0},z)$ \citep{Eisenhauer03}.

2290: Note that disc star counts must be differentiated three times, once

2291: to obtain $\nu(z)$ and twice in (5.3), strongly amplifying observational

2292: errors. \citet{Oort32} obtained $\rho(R_{0},z=0)\approx0.15$ \Ms\ 

2293: pc$^{-3}$. Hipparcos 65 years later has halved this

2294: to $0.076\pm0.015$ \Ms\  pc$^{-3}$ \citep{Creze98}.

2295: 

2296: Most halo stellar remnants are old white dwarfs that have cooled below

2297: 4000 K.

2298: Counts within 1 kpc height yield a halo stellar mass 

2299: density $\rho_{*}\sim0.044$ \Ms\ pc$^{-3}$. 

2300: Deeper surveys are underway, using

2301: kinematic cuts to isolate halo and disc white dwarfs, and selecting for

2302: bluer colours caused by molecular hydrogen (H$_{2}$) absorption in

2303: the red; cool white dwarfs have no other absorption lines. 

2304: From 34

2305: with halo kinematics, \citet{Oppenheimer01} conclude that such

2306: remnants comprise 2\% of the halo mass locally.

2307: However, \citet{Reid01} argue

2308: that this sample mixes at least two kinematical populations, with

2309: 20\% of `halo' white dwarfs actually in the thick disc (section \ref{sub:Chemo-dynamical-clues-in}).

2310: Because their disc density declines off the plane, the 

2311: true white dwarf halo population

2312: would be too small to account for much DM. This and the MACHO results

2313: exclude a maximal baryon thick disc at least locally.

2314: 

2315: The disc stellar \textit{column} density within a few scale heights

2316: of the plane is better determined because it requires \textit{only}

2317: a double differentiation of the star counts\begin{equation}

2318: \Sigma(<z\,\rm{kpc})=-\frac{1}{2\pi G\nu}\frac{\partial(\nu\overline{v_{\rm{z}}^{2}})}{\partial z};\end{equation}

2319: \citet{Oort32} found $\sigma(0.7)\sim90\,\rm{M}_{\odot}$ pc$^{-2}$; \citet{Creze98}

2320: review the methodology and efforts to refine his limit. \citet{Gould97}

2321: use deep HST images to count common disc M-dwarf stars far

2322: above the disc plane, normalize these to the locally determined density,

2323: and obtain $\sigma(1)=26\pm4$ \Ms\  pc$^{-2}$. \citet{Ollig01}

2324: obtain $\sigma(1.1)=35\pm10$ \Ms\  pc$^{-2}$ and show that the

2325: main uncertainty comes from the observationally restricted height

2326: into the halo; reliable tracers are sparse, blended with the populations

2327: of the discs, and cannot reach the several kpc height required to

2328: constrain tightly the DM halo.

2329: 

2330: Only beyond 30 kpc do we recover DM tracers, now dwarf satellites,

2331: GCs, and a smoothly distributed population of evolved RR Lyrae stars

2332: beyond 60 kpc \citep{Ivezic04}, their kinematics all implying DM

2333: \citep{Mateo98}. Examining these tracers, \citet{Ollig01} use mass

2334: models to conclude that the density toward the centre of the MWg's

2335: DM halo is uncertain by a factor of 1000. The relative contributions

2336: of stars and DM to the force above the disc plane cannot be separated

2337: easily, a `vertical disc-halo conspiracy'.

2338: 

2339: 

2340: \paragraph{Gas surface density}

2341: 

2342: The MWg RC $\Theta(r)$ is constrained by HI data to $\sim10$ \kms\

2343: internal accuracy. However, \citet{Lockman02} uncovers contradictions

2344: in the patchwork of datasets that comprise this curve in textbooks.

2345: He cautions us not to take it too seriously, especially between $1-2R_{0}$

2346: where its very uncertain gradient yields a 100\% uncertain HI surface

2347: density. The curve yields $v_{\rm{c}}^{2}/2\pi GR_{0}\sim226$ \Ms\ 

2348: pc$^{-2}$, $\sim2.5$ times that derived from luminous material within

2349: 700 pc of the disc \citep{Ollig98}. Most DM must be beyond this radius,

2350: in the halo. The IAU `consensus' rotational speed

2351: at the solar radius, $\Theta_{0}\equiv\Theta(R_{0}=8.0\pm0.4\,\rm{kpc})=220$

2352: \kms, exceeds slightly \citet{Ollig01}'s best estimate $200\pm10$ \kms.

2353: 

2354: The observed flaring of the HI layer beyond $2R_{0}$ is a constraint

2355: with different dependence on $R_{0}$ and $\Theta_{0}$. If the ISM 

2356: is pressurized only by thermally induced turbulent motions in the steady state,

2357: hydrostatic equilibrium applies, so \begin{equation}

2358: \frac{d\sigma_{\rm{g}}^{2}\rho_{\rm{g}}(z)}{dz}=\rho_{\rm{g}}(z)K_{\rm{z}}(z)\end{equation}

2359: with gas velocity dispersion $\sigma_{\rm{g}}=9-10$ km~s$^{-1}$

2360: and mass density $\rho_{\rm{g}}(0)\sim0.042$ \Ms\  pc$^{-3}$

2361: \citep{Malhotra95}, and $K_{\rm{z}}(z)$ the vertical force per

2362: unit mass. \citeauthor{Ollig01} estimate that neglecting pressures

2363: in the local ISM from the magnetic field (which is frozen to the ionized

2364: ISM) and cosmic rays (which are intense only near SF regions) makes

2365: less than 10\% error in the thickness of the gas layer at $R_{0}$

2366: and less at $2R_{0}$ where there is little SF.

2367: 

2368: A plausible DM density distribution is an isothermal spheroid\begin{equation}

2369: \rho_{\rm{dm}}(r,z;q)=\rho_{0}(q)(\frac{R_{\rm{c}}^{2}(q)}{R_{\rm{c}}^{2}(q)+r^{2}+(z/q)^{2}})\end{equation}

2370: with $R_{\rm{c}}$ the dark halo's core radius, $q$ the flattening

2371: from the $q=1$ sphere, and central density $\rho_{0}$ absorbs all

2372: dependence on $q$. With $R_{0}$ in kpc and $\Theta_{0}$ in \kms.

2373: \citet{Ollig01} find that\begin{equation}

2374: \frac{\rho_{\rm{dm}}(R_{0},\Theta_{0})}{10^{-3}\,\rm{M}_{\odot}\rm{pc}^{-3}}=\frac{11.5+3.8\times(R_{0}-7.8)\pm2}{q(26.7R_{0}/\Theta_{0})^{2}}\end{equation}

2375: The non-linear 

2376: coupling between these parameters is illustrated in Figure \ref{fig:ollingDMfig}.

2377: 

2378: 

2379: \begin{figure}[H]

2380: \begin{centering}\includegraphics[scale=2]{figure11.pdf}\par\end{centering}

2381: \caption{\label{fig:ollingDMfig}(Simplified from Figure 3 of \citealt{Ollig01}.)

2382: How constraints on MWg scaling constants relate to the gaseous velocity

2383: dispersion $\sigma_{g}$ and DM halo flattening $q$. Each panel chooses

2384: a $\Theta_{0}/R_{0}$, together spanning their uncertainties;

2385: numbers along the top of each panel give the local rotation speed $\Theta_{0}(R_{o})$

2386: in \kms. Gray shading spans current estimates of $R_{0}$ as obtained

2387: from parallax measurements of the Galactic Centre \citep{Eisenhauer03}.

2388: The vertical axis shows a large range over $\sigma_{g}$, with the

2389: two solid (\full) horizontal lines at \citet{Malhotra95}'s

2390: bounds. Hatched shading spans uncertainties in the column density

2391: of the stellar disc. The dark dashed (\broken) curve shows the

2392: spherical boundary between oblate and prolate shapes of the DM halo

2393: and the lighter dashed (\dashed) curves show the range. }

2394: \end{figure}

2395: 

2396: 

2397: \paragraph{Implications for the DM halo of the MWg}

2398: 

2399: Motions of the Sagittarius dwarf galaxy permit either a prolate or spherical 

2400: DM halo for the MWg \citep{Ibata01a}. However, a `Field of Streams' discovered

2401: in the SDSS by \citet{Belokurov06} shows that the star stream

2402: from this galaxy wraps around the MWg two or three times. \citet{Fellhauer06}

2403: argue that this coherence requires a close to spherical inner DM halo.

2404: (Section \ref{sub:Chemo-dynamical-clues-in} discusses MWg star streams.)

2405: \citet{Ollig01} use \citet{Malhotra95}'s data on the vertical velocity

2406: dispersion of HI in the Solar Neighbourhood to constrain flattening

2407: of the DM halo; for allowable limits on $R_{0}$ and on $\Omega(R_{0})$,

2408: Figure \ref{fig:ollingDMfig} shows that the MWg halo is spherical

2409: or barely oblate and has density in the Solar Neighbourhood $(11\pm5)\times10^{-3}$ \Ms\  

2410: pc$^{-3}$ (0.42 GeV/c$^{2}$ cm$^{-3}$). Note that this is

2411: within the $\pm0.015$ \Ms\  pc$^{-3}$ uncertainty of the stellar

2412: density from Hipparcos as mentioned above \citep{Creze98}, avoiding

2413: substantial DM in the local disc.

2414: 

2415: 

2416: \subsection{Modes of ongoing SF}

2417: 

2418: One cannot hope to model the evolution of galaxy baryons without understanding

2419: SF quantitatively. Unfortunately, dust obscures most SF at visible-light wavelengths.

2420: Stars form within

2421: an active, multi-phase ISM whose dynamics are only starting to become

2422: tractable as long-IR and sub-mm detector arrays have grown in number of

2423: pixels, and as

2424: 3D hydrodynamical simulations have become more realistic. It

2425: now appears that stars tend to form today in the

2426: coldest, molecular phase in galaxy discs. SF is stimulated above average

2427: levels by any non-axisymmetric instability that sweeps through the

2428: disc, usually triggered by close tidal passage or merger of a

2429: neighbouring galaxy. This has been quantified observationally \citep[for example][]{Larson78},

2430: and simulated in detail \citep{Mihos96}. Examining SF

2431: is beyond our scope, see \citet{Evans99} for example.

2432: Here we concentrate on its global aspects that are crucial

2433: to modelling galaxy evolution.

2434: 

2435: \subsubsection{\label{sub:Quantification-of-ongoing}Quantification of ongoing SF.}

2436: 

2437: How can SF rates be measured accurately in different types of galaxies and environments?

2438: SF manifests mainly through the high luminosities of the most massive

2439: MS stars formed. So, its unique signature is brief

2440: (few Myr), intense UV ionizing radiation from hot stars that is

2441: absorbed by gas in the star forming complex. Such HII

2442: regions emit hydrogen recombination lines in the visible-band

2443: (Balmer series), UV (Lyman), and IR (Paschen and Brackett), forbidden (i.e., electrical non-dipole

2444: transition) spectral lines of partially ionized oxygen and other elements,

2445: strong UV fluxes, free-free

2446: radiation in the radio continuum, and synchrotron radiation from energetic

2447: particles produced in SNe \citep[for example][]{Dopita05}. 

2448: In fact, because stars form in dusty GMCs, 

2449: graphitic and silicaceous grains quench UV and visible radiation, and

2450: recycles much radiation into IR and mm wavebands. 

2451: The signatures of SF at shorter wavelengths

2452: are easily extinguished. Thus IR photometry from space of heated dust

2453: is most effective at measuring the true SF rate. Abundances in the gas-phase

2454: are set both by astrophysical processes within stars and by chemistry

2455: on the surfaces of grains and gas molecules. 

2456: 

2457: The SF rate is set by the energetic photons

2458: (and particles) emitted by the most massive stars.

2459: Determining the total mass of a star burst therefore

2460: requires very uncertain extrapolation of the IMF to the lower masses

2461: that constitute most of the event. We encountered this problem in

2462: section \ref{sub:Evidence-from-other} when constraining stellar $\Upsilon_{d}$.

2463: What is observed as ongoing SF is the integrated result of

2464: many HII regions of diverse age, mass, size, pressure, and chemical abundance.

2465: K-band luminosity measures the integrated light of

2466: older stars  to establish 

2467: the efficiency of SF across a galaxy\begin{equation}

2468: b=\frac{\rm{SF~rate}_{\rm{present}}}{\left\langle \rm{SF~rate}_{\rm{past}}\right\rangle }\end{equation}

2469: (for example $\sim5\%$ in isolated dwarf galaxies), with

2470: average past SF rate $\left\langle \rm{SF~rate}_{\rm{past}}\right\rangle $.

2471: H$\alpha$ flux can be converted into the total flux of ionizing photons

2472: using recombination line transition probabilities, determining the

2473: rate of forming massive stars; by assuming a form of the IMF, $\rm{SF~rate}_{\rm{present}}$

2474: can be inferred \citep[for example][]{Kennicutt94}. More readily

2475: observable {[}O~II]$\lambda\lambda$3726,3729 is often a surrogate

2476: for H$\alpha$ in studies of high-redshift galaxies, but is less certain

2477: because of the strong temperature dependence of collisionally excited

2478: forbidden lines. \citet{Kennicutt98} reviews determinations of SF rate;

2479: he shows that the present SF rate characterized by parameter $b$ varies

2480: strongly with Hubble type: $b$ is often $<0.07$ for Sa discs, 0.3

2481: for Sb discs, and 1 for Sc discs, all with large dispersions. If the

2482: SF rate is taken to decline exponentially, then the e-fold timescale

2483: is only $\sim2$ Gyr for Sa's, rising to $\sim5$ Gyr for Sb's.

2484: 

2485: Can global SF rates be parametrized for semi-analytic models (and for

2486: mixed N-body and gas dynamical models) of galaxy evolution?

2487: Are such prescriptions sensitive to environment? For example, tides

2488: between galaxies in rich clusters occur more often than in low-density

2489: environments. But such encounters at the large velocity dispersion

2490: of the cluster will be `impulsive', with relatively small

2491: effect compared to the slow, nearly parabolic encounters of galaxies

2492: in groups. Thus, there is a complex trade-off between destabilization

2493: from frequent impulses (`harassment', Moore

2494: \etal 1996), and the effects of slow encounters, as modelled for

2495: example by \citet{Mihos96}.

2496: 

2497: Galaxies undergo major bursts of SF,

2498: often triggered by tides from a passing companion. A galaxy is

2499: star bursting mildly if $b=2-3$ and strongly if $b>10$.

2500: Bursts trigger on a dynamical collapse time: a 10 kpc radius with

2501: $10^{10}$ \Ms\  collapses in 200 Myr; mergers can speed this up.

2502: $b$ is measured most effectively by combining SDSS visible-band and

2503: \textit{GALEX} space-UV photometry; thus \citet{Salim05} find that

2504: $\sim20\%$ of galaxies with $\sim10^{8}$ \Ms\  to $\sim5\%$ of

2505: those with $\sim10^{11}$ \Ms\  have had a significant star burst

2506: within the last Gyr. Similarly, \citet{Heckman05} have studied 74

2507: nearby galaxies with high far-UV luminosities ($3-30$ \Ms\  yr$^{-1}$

2508: converting into stars) and find that galaxy mass and UV SB correlate

2509: inversely. Using the SDSS, \citet{Brinchmann04} find that in the

2510: local universe most stars form in Sgs that exceed $10^{10}$ \Ms\ 

2511: ($\sim15\%$ being AGN) in regions $100-500$ pc across, with $\sim20\%$

2512: in mild star bursts and 3\% in intense ones $(b>10)$.

2513: 

2514: `Passive' SF in high SB galaxies occurs within GMC cores at densities

2515: $10^{2}-10^{3}$ cm$^{-3}$; section \ref{sub:Low-surface-brightness}

2516: discusses SF in LSBg's. Young stars are dust enshrouded for their

2517: first 1-2 Myr, and radiate most ionizing UV within the first $\sim6$

2518: Myr, thereafter photo-dissociating their molecular envelopes over

2519: $\sim10$ Myr. For the next $50-100$ Myr, burst luminosity

2520: increases by only 10\% \citep{Leitherer99} as GMC destruction

2521: ends SF.

2522: 

2523: Broad-band colours of visible light age SF crudely \citep{Whitmore95}.

2524: Sensitivity to SF within the last 100 Myr is much improved by adding

2525: the \textit{GALEX} NUV and FUV filter bands (Figure \ref{fig:Spectral-evolution-of}a).

2526: More accurate timing of an ongoing or recently ended burst follows

2527: from details of the gaseous emission-line flux ratios. \citet{Gonzalez99}

2528: show that the He~I/H$\beta$ ratio varies with the age of the

2529: burst: $0.10-0.12$ until 4 Myr, then declines steeply below $0.05$.

2530: The {[}O~III]$\lambda$5007/H$\beta$ ratio starts near 4.5, declines

2531: to $\sim1.2$ by 2.5 Myr, and reaches 0.5 by 3 Myr. Near-IR images of

2532: hydrogen and helium emission can map SF in dust shrouded discs. In

2533: the MWg, emission from transitions between barely bound atomic levels

2534: of hydrogen and helium are mapped with radio interferometer arrays

2535: across very extinguished parts of GMCs \citep[for example][]{Gordon03}.

2536: 

2537: \subsubsection{Sites of SF and dependence on gas density.}

2538: 

2539: Where do stars form? How does gas surface density

2540: set the disc SF rate \citep{Schmidt65}? 

2541: How the rate of `passive' SF depends on the surface density

2542: of hydrogen $\sigma_{\rm{gas }}$, and the threshold above which

2543: stars can form, was estimated empirically by \citet{Kennicutt98}

2544: from the discs of small-bulge Sgs as\begin{equation}

2545: \sigma_{\rm{SF}}\propto\sigma_{\rm{gas}}^{1.4}\end{equation}

2546: His value agrees with the theoretical `disc instability' criterion

2547: for cloud collisions \citep{Toomre84}, although both \citet{Schaye04}

2548: and \citet{Noordermeer06} have subsequently mapped ongoing SF in

2549: large bulge high SB galaxies below half this value. \citet{Silk77} gives another

2550: estimate\begin{equation}

2551: \sigma_{\rm{SF}}\propto\sigma_{\rm{gas}}\Omega_{\rm{gas}}\end{equation}

2552: with $\Omega_{\rm{gas}}$ the orbital speed of the gas clouds. \citet{Thilker05}

2553: have used \textit{GALEX} to find that stars are forming in some galaxies

2554: at $2-4$ times the radius of the visible-band disc with $\sigma_{\rm{gas}}$

2555: below Kennicutt's thresholds as established at smaller radii. $\rm{SF~rate}_{\rm{present}}$

2556: depends on the mass fraction in dense gas, whose value is set by turbulence

2557: \citep{Elmegreen02}. \citet{Kennicutt98} finds\begin{equation}

2558: \rm{SF~rate}_{\rm{present}}\,\rm{(M}_{\odot}\rm{\, yr}^{-1})=\frac{L(H\alpha)}{1.26\times10^{34}\rm{\, W}}\end{equation}

2559: although this misses the, perhaps majority, SF that is extinguished

2560: by dust.

2561: 

2562: Until SF ends with the conversion of all ISM into a warm phase, it can propagate over

2563: a $\lesssim1$ kpc scale \citep{Zhang01};

2564: this behaviour is evident in mergers and also as a clear sequence

2565: of `super-shells' in the LMC where up to half of the stars form

2566: adjacent to such high-pressure HII regions (true also in the MWg).

2567: 

2568: 

2569: \subsubsection{\label{sub:Raw-material----}Raw material --- effects of the state

2570: of the ISM.}

2571: 

2572: A major uncertainty in the physical state of the ISM and

2573: also the spectrum from hot stars that bathes HII regions is

2574: the depletion of elements from the gas phase onto

2575: dust grains and polycyclic aromatic hydrocarbon (PAH, a planar-hexagonal

2576: carbon lattice built from benzene rings) molecules formed in dusty

2577: bubbles around AGB and RGB carbon-rich stars; the size distribution

2578: of grains (power-law up to large sizes, declining exponential thereafter);

2579: grain evaporation by thermal and perhaps chemical sputtering, grain

2580: shattering, and shock heating from SN; the photo-dissociation of H$_2$ into HII;

2581: and PAH photo-dissociation

2582: in HII regions. Chemical abundances derived from emission lines

2583: show how metals still in the gas phase vary across unobscured HII

2584: regions. Most otherwise abundant elements, originally as gas, have

2585: depleted onto dust grains; sulphur is one exception, but its emission

2586: intensity depends on its temperature, abundance, and level of ionization.

2587: 

2588: The ISM of the MWg has been reviewed lucidly by \citet{Ferriere01}.

2589: Despite $\sim1000$ generations of SF in the MWg disc each of $\sim10$

2590: Myr duration \citep{Bland-Hawthorn04}, $\sim10^{7}$ \Ms\  of dust

2591: and $4\times10^{9}$ \Ms\  of gas remain to extend SF to 30 kpc radius.

2592: Gas subdivides into a cool layer of $\lesssim0.3$ kpc scale height,

2593: and a warm neutral (Lockman) and ionized (Reynolds)

2594: layer to 0.6 kpc. Half of its mass is molecular hydrogen

2595: (H$_{2}$) within cores of GMCs of mass $(0.1-2)\times10^{6}$ \Ms. 

2596: Velocity crowding of multiple clouds blurs our view. Hence,

2597: over the last decade, less edge-on Local Group galaxies mapped in surveys with

2598: the BIMA array (N. hemisphere) and 4-meter aperture NANTEN dish (S.

2599: hemisphere) have provided our best views. \citet{Blitz06} review what

2600: these surveys have revealed about the molecular

2601: ISM, using emission from the CO molecule as a tracer for the molecular

2602: gas. This gas is correlated with HI in that every GMC is found on

2603: an HI filament but there are many filaments without CO. CO concentrates

2604: toward galaxy centres, with exponential scale lengths comparable to

2605: that of the stellar disc. To form stars, HI must clump to become

2606: a filamentary GMC, increasing its surface density up to twenty-fold;

2607: this conversion is almost complete in high-pressure, HI deficient

2608: regions such as galactic nuclei.

2609: 

2610: HII regions \citep{Oey97} and wind bubbles from massive stars break

2611: out of the GMC along its density gradient, expanding as the star burst

2612: progresses until the gas pressure balances with the lower-density,

2613: diffuse ISM. In balance, bubbles then coast in a momentum-conserving

2614: state (in which they pass most of their time) that is set by the density

2615: of the ISM phase with the largest volume filling factor, until the

2616: swept-up shell of ISM fragments by turbulence. In this way and with

2617: SN, the GMC core dissolves $\sim1$ Myr after SF is triggered \citep{Dopita06}.

2618: A sequence of cluster ages and correlated CO emission in the LMC shows

2619: that a cloud dissolves in $\sim20-25$ Myr \citep[and references therein]{Blitz06};

2620: HII regions then pass $\sim\frac{1}{4}$ of their lives in a

2621: quiescent phase.

2622: 

2623: The other half of the ISM mass in the MWg is HI in a complex \citep{Cox05},

2624: multiphase topology that is too uncertain to model

2625: evolving SF \citep[although, see the daring simulations of][]{deAvillez00}.

2626: Instead, SF is usually modelled in a uniform pressure and density

2627: ISM \citep[for example][and references therein]{Dopita05}. In the

2628: disc, HI is often optically thick and the uncertain transverse component of its

2629: non-circular motions makes uncertain

2630: $\sigma_{\rm{gas}}$ across much of the MWg (section \ref{sub:Evidence-in-the}).

2631: Cosmic rays, heavy

2632: nuclei accelerated at the shock fronts of expanding supernova bubbles \citep{Ferriere01},

2633: deposit a third of the energy into the ISM.

2634: 

2635: Once more than $\sim30$

2636: successive SNe have detonated, their individual `chimneys' (in

2637: simulations like \citealt{deAvillez00} rising 500 pc above the disc)

2638: merge into a `superbubble' \citep[see the review of][]{Veilleux05}.

2639: These x-ray emitters (thermal spectrum, typically $\sim10^{5-6}$ \Ms)

2640: are surrounded by H$\alpha$-emitting shells up to $\sim1.5$

2641: kpc diameter; the LMC has particularly clear examples. If either structure reaches high enough,

2642: its thermalized gas redevelops momentum

2643: down the pressure gradient, and the flow can entrain and `break out'

2644: of the disc as a `galactic fountain': once above $\sim$250

2645: pc altitude, it becomes Rayleigh-Taylor unstable from declining ambient

2646: densities and eventually drips as descending

2647: sheets that pump entrained ISM (including

2648: metals, for example \citealt{Kenney99}) across a galaxy to

2649: smear out primordial abundance gradients. 

2650: Indeed, at $<1$ kpc \citet{Lockman03} finds dense HI clouds

2651: that are too low mass to self-bind gravitationally; they have disc kinematics

2652: so were presumably uplifted by correlated SNe and the resulting `fountain'.

2653: \citet{Pidopryhora06} map a superb example $\sim7$ kpc distant

2654: that reaches $\sim1$ kpc into the MWg halo, contains $\sim10^{6}$ \Ms\  

2655: of hydrogen, and has total energy $\sim10^{53}$ erg.

2656: 

2657: ISM porosity to supernova blast waves sets the lateral extent

2658: of a star burst, the chemical yield, and how efficiently gas is ejected

2659: from a potential (see section \ref{sub:Feedback}).

2660: The other important but poorly observed parameter \citep[for example][]{Veilleux05}

2661: is the thermalization efficiency $\epsilon$ --- the fraction of mechanical

2662: energy from all SNe not yet radiated before their blast waves merge;

2663: it is degraded into turbulence thence entropic heating. $\epsilon$

2664: depends on the pressures and scale heights of

2665: the various ISM phases, and on the clustering of SNe. Estimated values

2666: in star bursts vary widely because of uncertain ISM properties, from

2667: $\epsilon=0.005$ (appropriate for a cloudy ISM with many GMCs and

2668: a mix of SNe Ia and II) to $\epsilon=1$ (appropriate for a galaxy

2669: bulge with only SNe Ia). ISM has been incorporated into semi-analytic

2670: models of SF by scaling parameters from a three-phase description

2671: \citep[see the review by][]{Cox05}. In this picture, a poorly constrained

2672: topology of static cool clouds, warm interface, and pervasive hot

2673: phase from intersecting supernova remnants forms and evolves toward $\epsilon=1.$

2674: Blast waves evaporate clouds, and the resulting warm ablata balance

2675: cooling from denser regions.

2676: 

2677: \subsubsection{\label{sub:Low-surface-brightness}SF in low-SB galaxies.}

2678: 

2679: Selections inhibit discovery of LSBg's, so the correlations discussed

2680: previously are only now being extended as targeted searches and surveys

2681: like the SDSS uncover examples over the full range of galaxy mass,

2682: in voids, and on the rims of matter filaments. Most LSBg's have

2683: no bulge or bar and little dust.

2684: Their blue colours (young stars hence slow, continuous SF), 

2685: metallicities ($0.05-0.5$ solar, a few $>1$), and

2686: gas content are less evolved than in high SB galaxies; they are not faded.

2687: They are often gas rich yet have HI

2688: densities below \citet{Kennicutt98}'s threshold for SF (for example

2689: Figure \ref{fig:Example-LSBg-rotation}), and sparse CO suggesting

2690: no GMCs. \citet{Mihos99} model their ISM and conclude that H$_{2}$

2691: can be substantial within a few kpc radius but its warmth

2692: $\sim30-50$ K inhibits SF. 

2693: They have extreme HI masses for their blue luminosity, up to 10 times

2694: those of high SB galaxies. 

2695: 

2696: \begin{figure}

2697: \begin{centering}\includegraphics[scale=1.55]{figure12.pdf}\par\end{centering}

2698: \caption{\label{fig:Example-LSBg-rotation}RC (vertical lines)

2699: and HI surface density (\fullcircle) of LSBg

2700: UGC 6614. Only at 30 kpc does the surface density rise to meet

2701: the critical value of \citet{Kennicutt98} (\dotted) to form stars.

2702: This figure

2703: originally appeared in the Publications of the Astronomical Society of the

2704: Pacific \citep[PASP, 109, 745]{Bothun97}.  Copyright 1997,

2705: Astronomical Society of the Pacific; reproduced with permission of the

2706: Editors.}

2707: \end{figure}

2708: 

2709: \subsubsection{\label{sub:Jet-and-radio-lobe}Jet and radio-lobe induced SF.}

2710: 

2711: Jets can trigger some SF by crushing ISM through direct impact or thermal shocks

2712: driven by the expansion of the cocoon/radio lobe \citep[and references therein]{Bicknell00,Fragile04,Odea04},

2713: to explain aligned radio continuum jets and SF

2714: (for example Minkowski's Object near NGC 541, \citealt{Croft06}).

2715: However, in section \ref{sub:Feedback} we discuss suppression of SF (`feedback') by jets.

2716: \\

2717: 

2718: In summary, disc baryons provide our best opportunity to trace motions

2719: in DM dominated regions. 

2720: We have shown how to measure the SF rate and how it depends on the

2721: densities of the diffuse, multi-phase ISM. 

2722: The diversity of SF complicates interpretation:

2723: uncertainties about dust content and IMF make $\Upsilon$ uncertain, complicating isolation of the

2724: DM component.

2725: 

2726: \section{\label{sec:Matter-Transfer-Within}Feedback}

2727: 

2728: To parametrize

2729: models of the global evolution of star-forming galaxies, the key question

2730: is to what extent does energy injected into the surrounding

2731: ISM (through UV radiation and winds/SNe from massive stars, production

2732: of heavy elements that alter the gas cooling rate, etc) either promote

2733: or discourage continuing SF, and redistribute the remaining ISM among

2734: its phases? The astrophysical processes responsible are denoted collectively

2735: \citep{White78} \textbf{feedback}. Because it acts within the underlying

2736: galaxy DM potential well, feedback alters the dependences of the SF and chemical evolution rates

2737: on galaxy mass.

2738: 

2739: \begin{figure}

2740: \begin{centering}\includegraphics[scale=0.28]{figure13.pdf}\par\end{centering}

2741: \caption{\label{fig:Data-for-the}\citep[From][used with permission]{Benson03}. The

2742: galaxy luminosity function $\Phi$:

2743: the number of galaxies per million pc$^3$ volume and per

2744: magnitude in apparent brightness versus the $-\log$ luminosity of a galaxy (in

2745: units of its absolute magnitude, the MWg being $-20.7\pm0.2$).  Data

2746: are \opencircle, \opensquare,

2747: and stars with error bars. Shown are models of the DM halo-mass function

2748: without feedback, (diagonal \broken), and with various feedbacks

2749: mentioned in the text as curves (cooling (\dotted); photoionization

2750: (\longbroken); merging (\full)).}

2751: \end{figure}

2752: 

2753: The crucial r\^ole of feedback in galaxy evolution appears when

2754: modelling the observed luminosity function (LF) of galaxies. In the

2755: $\Lambda$-CDM scenario, the redshift dependence of the hierarchical assembly

2756: of DM halos can be modelled straightforwardly because only gravity operates.

2757: The DM halo mass distribution is predicted either from

2758: N-body simulations \citep[for example][]{Springel05} or through the

2759: analytical framework of the extended Press-Schechter theory \citep{Press74,Somerville99,Cole00}.

2760: Because DM halos build by amplifying and merging an initial Gaussian

2761: distribution of density fluctuations with a roughly scale-invariant

2762: power spectrum that has grown into the non-linear regime, the resulting mass

2763: distribution has an exponential high-end

2764: and a roughly power-law low-end, as observed \citep{Schechter74}.

2765: However, if one assumes that CDM is modified

2766: by a single $\Upsilon$ to yield the baryonic mass distribution (but

2767: recall in section \ref{sub:Kinematics:-the-fundamental} the tight $\Upsilon-\sigma_{\rm{e}}$

2768: correlation in Egs), one finds that model and observed LF's disagree.

2769: This model LF (\broken\ in Figure \ref{fig:Data-for-the})

2770: has two problems: its low-luminosity

2771: slope over-predicts faint galaxies, and it declines steeply at much

2772: higher luminosity (mass) than observed. Adding a simple prescription

2773: for gas cooling (\dotted) further over-predicts the faint end

2774: because lower mass halos are cooler virially. However, cooling helps

2775: at the bright end because gas cools slowly in the

2776: massive, virially hot DM halos. Low masses are further suppressed

2777: by re-ionization after the first stars and AGN form

2778: (\longbroken). Adding mergers of small galaxies into larger ones (\full) further boosts

2779: the luminous end.

2780: Conversely,

2781: this end can be reduced by suppressing cooling in massive Egs with AGN via reheating shocks

2782: from their jet-emitting supermassive black holes \citep{Croton06,Bower06}, see Appendix B.

2783:  

2784: While matching the observed galaxy LF, a successful model of galaxy

2785: evolution must also reproduce the observed K-band TFR. The

2786: zeropoint and slope of the TFR depends on what fraction of baryons

2787: cool to form K-luminous stars, while their scale length in

2788: the DM halo modifies the RC. Moreover, the $\Lambda$-CDM scenario

2789: yields a continuum of SF rates in galaxies because it does not

2790: inhibit accretion of cold gas that would equip all galaxies with a

2791: reservoir from which to form stars today. Yet, the SDSS has quantified

2792: (Figure \ref{fig:Galaxy-color-distribution}) earlier impressions

2793: \citep{Kauffman03} that galaxies divide bimodally

2794: by colour and mass into a blue sequence of active, SF  galaxies and a sequence of 

2795: `red and dead' early-types that appeared after $z=1$ (for example, \citealp{Wake06})

2796: with AGN-like emission spectra \citep{Cooper06}.

2797: 

2798: \begin{figure}

2799: \begin{centering}\includegraphics[scale=0.4]{figure14.pdf}\par\end{centering}

2800: \caption{\label{fig:Galaxy-color-distribution}Galaxy colour distribution

2801: from SDSS \citep[used with permission]{Baldry04}. 

2802: Left: data corrected for incompleteness, with contours showing the number

2803: of galaxies. Right: deconvolution into

2804: `red and dead' (top ($\full$)) and star forming (bottom ($\broken$))

2805: components.

2806: Separation is even cleaner when \textit{GALEX} colours are used too.}

2807: \end{figure}

2808: 

2809: \begin{figure}

2810: \begin{centering}\includegraphics[scale=0.57]{figure15.pdf}\par\end{centering}

2811: \caption{\label{fig:Feedback-flow-chart.}The baryon flows that operate in

2812: stellar and AGN feedbacks, and their outcomes. 

2813: The second box from the top is treated currently with numerical

2814: simulations because observational constraints are rudimentary. The other steps are discussed

2815: in the text.

2816: The flows operating in disc starbursts are highlighted at right.}

2817: \end{figure}

2818: 

2819: In summary, feedback and other baryonic gas dynamics alter galaxy

2820: properties drastically; they `quench' subsequent SF

2821: and build up mass by inhibiting flows of energy and angular momentum.

2822: Multiple routes in Figure \ref{fig:Feedback-flow-chart.} 

2823: build the red galaxy sequence from the blue;

2824: the effectiveness of some depend strongly on environment so are detailed

2825: in section \ref{sec:Interactions-with-Other}. \citet{Baugh06} and \citet{Cole00}

2826: review comprehensively the parameters of galaxy evolution models.

2827: In the following, we merely outline the most important processes.

2828: 

2829: \subsection{Hot and cold modes of gas accretion}

2830: 

2831: Most accretion occured at high redshift, in two modes: 1) `hot'

2832: gas falls into the DM halo at its virial temperature, heating

2833: faster than the cooling time in an almost spherical shock.

2834: 2) `cold' merging gas cools faster

2835: than it is shock heated so collapses first into and then perhaps along

2836: a filament \citep{Keres05}. In the present epoch, the cold mode is

2837: located by simulations \citep{Keres05} at $<1.8\times10^{11}$ \Ms\ 

2838: because of lower infall velocities. 

2839: In contrast, the simulations by \citet{Cox04} highlight how the final merger

2840: can shift gas from a cold, SF phase to a hot, non-SF one whose ionization

2841: depends on the metagalactic UV radiation field, and the uncertain leakage

2842: of ionizing photons from the disc and from halo planetary nebulae.

2843: 

2844: Unclear is how a cold accretor sheds its gravitational energy to bind

2845: to a galaxy: through a strong shock near the disc, or gradually enough

2846: for pressure gradients to settle gas gently into disc rotation? 

2847: \citet{Fraternali06}

2848: consider the extended HI gaseous halos of disc galaxies \citep[for example][]{Swaters97}, and 

2849: conclude from the failure of velocity models to avoid an HI outflow

2850: that the galaxy must be accreting from the intergalactic medium.

2851: A massive yet

2852: isolated HI cloud observed by \citet{Minchin05} cannot be tidal

2853: debris \citep{Bekki05}. Will it become a cold accretion? Metallicity

2854: measurements can discriminate provided that ionized gas is present

2855: or if there is a background quasar spectrum to absorb its light with

2856: neutral gas. Other clues come from the orbits of Local Group metal-poor globular clusters

2857: and other satellites of the MWg and M31 \citep[including M33,][]{Koch05},

2858: which are flattened along the major axis of their velocity ellipsoids

2859: to indicate non-spherical potentials. No comparable constraints exist

2860: for hot accretion at high redshifts, probably because its spherical

2861: distribution would be too diffuse to detect with present day x-ray

2862: telescopes.

2863: 

2864: 

2865: \subsection{\label{sub:Feedback}Mechanisms and scales of feedback}

2866: 

2867: `Classical' feedback in a forming, early-type galaxy arises from a large-scale

2868: wind established in the ISM by overlapping hot, blast-wave bubbles

2869: of supernova explosions (each $\sim10^{51}$ erg) and stellar winds (each $10\%$

2870: as strong). Recently identified as a major feedback are the

2871: supermassive black-holes in AGN that drive particle jets 

2872: (section \ref{sub:Jet-and-radio-lobe}), their

2873: shocked cocoons and associated thermal winds over multi-kpc scales

2874: and each potentially $\sim10^{47}$ erg~yr$^{-1}$ for up to $0.1$ Gyr. 

2875: \citet{Veilleux05} discuss these outflows. An outflow

2876: is prominent where directed motion thermalizes, and usually has peak

2877: contrast against unrelated emission at a characteristic energy. For

2878: example, gas decelerating from $v_{\rm{s}}$ \kms\ clumps

2879: and attains post-shock temperature $T_{\rm{ps}}=0.11v_{\rm{s}}/(100$

2880: \kms) $10^{6}$ K that radiates x-rays strongly and is extinguished by

2881: photoelectric absorption, not directly by dust.

2882: The shocked gas may be further ionized by the AGN's `ionization cone', 

2883: although the AGN's radiation field dilutes rapidly with distance.

2884: Note that outflows

2885: are not the only way to sweep gas from a galaxy; in section \ref{sec:Interactions-with-Other}

2886: we discuss `ram pressure' stripping as a galaxy passes near

2887: the core of a cluster.

2888: 

2889: The rudimentary treatment of ISM in most semi-analytic models propagates

2890: blasts too rapidly through the hot phase, so clouds are

2891: crushed quickly to overproduce stars. An ISM more in tune with modern

2892: understanding (section \ref{sub:Raw-material----}) reduces porosity and

2893: confines blast bubbles adiabatically to retain most of the supernova energy.

2894: For example, \citet{Springel03} use a two-phase ISM whose components

2895: exchange mass by condensing into cool clouds, by evaporating in blast

2896: waves, by supernova heating, and are replenished by enriching, energetic wind

2897: outflows from massive stars. With this ISM they do predict a quiescent cosmic 

2898: history of SF that is consistent with

2899: red massive, early-type galaxies (Figure \ref{fig:Galaxy-color-distribution})

2900: and blue, lower mass galaxies (below $10^{12}$ \Ms;

2901: \citealt{Cattaneo06}) that cool and form stars.

2902: 

2903: Gas is also reheated and ejected by AGN. In simulations, an AGN is

2904: fuelled by mergers that usually stir star orbits into a bar or

2905: other transient structure, which then torque and shock gas into infall.

2906: The efficiency of feedback from a supermassive black-hole is set by the nature of the

2907: shock propagating medium, and by the efficiency of gravitational energy

2908: release $\epsilon_{0.1}$ (units of 10\%)

2909: such that $E=2\times10^{61}M_{\rm{BH},8}\epsilon_{0.1}$ erg 

2910: (black-hole mass in $10^{8}$ \Ms). To remove $M_{\rm{g},11}$

2911: (units $10^{11}$ \Ms) from a region of velocity dispersion

2912: $\sigma_{200}$ (units 200 \kms) requires $E=4\times10^{58}M_{\rm{g},11}\sigma_{200}^{2}$

2913: erg. The main accelerator at small radii is radiation pressure $P_{\rm{rad}}/k\sim2\times10^{7}L_{46}r_{\rm{kpc}}^{-2}$

2914: for ionizing luminosity in units of $10^{46}$ erg and radius in kpc, respectively;

2915: the process is only effective within $\sim10$ pc from AGN found locally

2916: and $\sim1$ kpc in quasars. Coupling AGN ionizing photons to gas is greatly

2917: enhanced in dusty clouds \citep{Dopita02}.

2918: 

2919: \citet{Springel05} find that a rapid star burst or AGN outflow can

2920: end SF abruptly. \citet{Cattaneo07} find that intermittent jets are

2921: most effective because they do not burrow beyond small radii, hence

2922: deposit their energy in the densest gas. Such jets reproduce quantitatively

2923: the colour-luminosity,

2924: -environment, and -morphology relations of the SDSS dataset,

2925: and form quickly the red and dead population of

2926: massive early-type galaxies. AGN feedback therefore

2927: establishes successfully the bimodality of galaxy types in massive

2928: galaxies whose potential wells are too deep for supernova feedback to eject

2929: gas, see Figure \ref{fig:Galaxy-color-distribution}. If accreted

2930: gas settled in while a luminous AGN was powered down, the subsequent

2931: effects on it after power up could be strong. Dust grains would charge

2932: electrically and be expelled, the cold phase removed, the density

2933: of the diffuse phase reduced, and the porosity of the ISM to supernova increased

2934: from $\sim10^{-4}$ to 1 over $\sim8$ Myr, all by a galaxy-wide ionization

2935: front. 

2936: 

2937: 

2938: \subsubsection{\label{sub:Galaxy-scale-winds}Galaxy-scale winds.}

2939: 

2940: A galaxy-scale wind regulates SF by heating and entraining ISM into

2941: the halo or beyond. Cooling $10^{6-6.5}$ K gas is imaged by the Chandra

2942: x-ray Observatory (\textit{CXO}), the kinematics of gas at $\sim10^{5.5}$

2943: K are constrained by \textit{FUSE} spectra of OVI$\lambda\lambda1032+1038$

2944: emission at the interface between the cooler bulk cloud and the enveloping

2945: ISM \citep[see the simulations of][]{Marcolini05}, post-shock gas

2946: at $10^{4}$ K is mapped across outflows in multiple visible-band

2947: and near-IR emission lines that can diagnose gas conditions, mid-IR emission

2948: from hot dust is mapped by the Spitzer Space Telescope, 

2949: and HI and molecular lines are mapped with radio interferometers. 

2950: 

2951: M82 and NGC~3079 \citep{Cecil01}

2952: provide particularly clear, multi-frequency views of SN-driven, galaxy-winds,

2953: while \citet{Strickland04} use \textit{CXO} to survey diffuse soft

2954: x-ray emission in a sample of edge-on Sgs that spans the full range

2955: of SF activity. In thin, star bursting discs, superbubbles blow out

2956: vertically, quickly energizing at least the galaxy halo and perhaps

2957: the IGM (although, $\Lambda$-CDM says that the virial radius of the

2958: dark halo of a massive galaxy is $\sim250$ kpc and existing x-ray

2959: telescopes are too insensitive to track gas to such low SB). In

2960: the sharp images of \textit{CXO}, \citeauthor{Strickland04} isolate

2961: diffuse emission from point sources (mostly low-mass x-ray binaries) and

2962: find that its SB in edge-on Sgs

2963: correlates with the mean SF rate per unit area. Most of the

2964: clumped emission in `windy' galaxies turns out to be ambient ISM clouds that have been

2965: crushed into pancakes and accelerated by the wind. Depending on how

2966: well it resists Kelvin-Helmholtz and Rayleigh-Taylor fluid instabilities,

2967: the shredding cloud may approach the wind terminal velocity \citep{Marcolini05}.

2968: 

2969: 

2970: \subsubsection{\label{sub:Ubiquitous-AGN-and}Ubiquitous AGN and supermassive black-holes.}\

2971: 

2972: \textit{SWIFT} satellite hard x-ray images and HST/ground-based

2973: long-slit spectra have uncovered supermassive black-holes

2974: in essentially all nearby galaxies with bulges \citep[for example][]{Kormendy95,Ho97},

2975: including some starbursts. The region of gravitational influence

2976: is small (at 4 Mpc distance, 1 arcsec for $2.5\times10^{8}$ \Ms)

2977: and in only three cases can one distinguish unambiguously between

2978: a dense star cluster and a supermassive black-hole (the MWg, M31 and NGC~4258, see the review

2979: of \citealt{Merritt06}). A few masses can be obtained from the circular

2980: Keplerian motion of compact masers within GMCs that orbit at the outer

2981: edge of a nuclear disc (NGC~4258, NGC~4945, and Circinus and NGC~1068

2982: with the complication of severely warped circumnuclear discs), see

2983: the review of \citet{Lo05}. More model dependent techniques to bound

2984: masses in more distant galaxies are reverberation maps from correlated

2985: broad line-AGN continuum flux variations \citep{Peterson04}, or when

2986: the AGN gravitationally lenses a background galaxy into multiple images

2987: \citep{Rusin05a}. 

2988: 

2989: Our sharpest view of an AGN comes from studies of the supermassive black-hole in the

2990: MWg, Sag.~A{*}, which flares in the near-IR/x-ray periodically as it

2991: is fuelled from a relativistic orbit

2992: \citep{Schodel03}. From precise orbits of stars

2993: S02 ($15.3\pm0.34$ yr period) and S0-16 (comes within 600 Schwarzschild

2994: radii, 45 AU), \citet{Mouawad05} and \citet{Ghez04} derive mass

2995: $3.5\pm0.3\times10^{6}$ \Ms, and indeed will soon measure relativity

2996: effects such as perigalacticon advance. 

2997: 

2998: Masses of central black holes correlate with host galaxy properties \citep{Marconi03,Haring04,Ferrarese00,Gebhardt00}

2999: \begin{equation}

3000: M_{\rm{BH},8}=1.66\pm0.24(\sigma/200)^{4.86\pm0.43},\end{equation}

3001: (units of $10^{8}$ \Ms) implying that the black hole has $\sim0.3\%$ of a galaxy's

3002: baryon mass. There is some scatter, for example the $\sigma$'s

3003: of the MWg \citep{Tremaine02} and M87 \citep{Cappellari06} are too

3004: small for their correlation masses, whereas the supermassive black-hole masses in M31

3005: ($1.1\times10^{8}$ \Ms\  \citealt{Bender05}),

3006: M32 ($(2.5\pm0.5)10^{6}$ \Ms\  \citealt{Verolme02})

3007: and Cen~A \citep{Krajonovic06} are twice those predicted, so perhaps are powered

3008: down.

3009: Indeed,

3010: \citet{Revnivtsev04} suggest that the x-ray bright reflector at Sag. B2 $\sim100$

3011: pc away indicates that the MWg's 

3012: supermassive black-hole emitted $L\sim1.5\times10^{39}$ erg~s$^{-1}$

3013: (i.e., $10^{6}$ brighter than today but still only 1\% that of a

3014: typical AGN) in the band 2-200 keV for at least a decade 300-400 years

3015: ago.

3016: 

3017: \citet{Wyithe03} fit simultaneously

3018: the $M_{\rm{bh}}-\sigma_{\rm{gal}}$ relation and the evolution,

3019: shape, and zero-points of the quasar visible-band and x-ray luminosity

3020: functions out to $z=5.5$ in a model wherein a black hole powers up whenever

3021: it feasts on gas delivered by a major merger 

3022: \citep[for example the DM simulations of][]{Volonteri03}.

3023: They predict $M_{\rm{bh}}\propto V_{c}^{5}$, independent of redshift.

3024:  \citet{Rafferty06} find cavities in the intracluster medium that

3025: were inflated by large radio jets. AGN are able energetically to balance

3026: cooling in more than half of the 33 clusters studied. The most powerful

3027: example of this heat source was found by \citet{McNamara05} in a

3028: cluster at $z=0.22$: a radio source $\sim550$ kpc long of $\sim6\times10^{61}$ erg 

3029: has so heated all the intracluster gas over several Gyr that it

3030: has been unable to cool onto the central cD galaxy to form stars.

3031: 

3032: In summary, both supermassive black-hole and starburst winds are driving outflows

3033: that are currently extensive enough to impede mass buildup and to

3034: regulate the SF rate over the past Gyr.

3035: These and other energy exchange mechanisms in the ISM are sufficiently complex and 

3036: poorly understood that our treatments of feedback in galaxy evolution remain

3037: very uncertain.

3038: 

3039: 

3040: \section{\label{sec:Interactions-with-Other}{R\^ole

3041: of group/cluster environment in galaxy evolution}}

3042: 

3043: Section 6 outlined the evolutionary r\^oles of internally driven feedback.

3044: However, evolution can also be stimulated by

3045: events within the virialized, hence hot (${\sim10}^{8}$ K) environment of a

3046: galaxy cluster. 

3047: Ram pressure induced as a galaxy

3048: transits this medium can erode the galaxy ISM \citep{Gunn72}, and it and stars

3049: can be disrupted or merged by tides.

3050: We therefore address how environment affects evolution,

3051: first considering how cluster studies have constrained DM.

3052: 

3053: \subsection{Dark matter}

3054: 

3055: 

3056: \subsubsection{Evidence from galaxy groups.}

3057: 

3058: Evidence for DM around galaxies in groups and bound pairs is slim

3059: \citep{Persic96}. Mass comparable to that in the stars is estimated

3060: from x-ray images \citep[see the review by][]{Mulchaey00} by assuming

3061: that this gas is in hydrostatic equilibrium in the group's gravitational

3062: potential. This is reasonable for the subset of groups with regular,

3063: circular shape given the short sound-crossing time compared to local

3064: cooling time. The intragroup medium has a significant abundance of

3065: heavy elements, usually attributed to galaxy-scale winds (section \ref{sub:Galaxy-scale-winds}).

3066: But the relative

3067: effectiveness of supernova heating in the wind to that of the gravitational

3068: potential of the group is uncertain because x-ray images even with

3069: \textit{CXO} have not established the efficiency of energy transfer

3070: from SNe to the gas because of uncertain gaseous filling factors \citep{Strickland04}.

3071: The standard assumption is that ongoing galaxy winds are unimportant

3072: heat sources and that the hot gas is spherically symmetric. 

3073: assumptions on gas temperature and metallicity.

3074: If the radial x-ray SB follows a \citet{King62} profile, 

3075: temperature can also be determined at projected radius.

3076: x-ray telescopes before \textit{XMM/Newton} and \textit{CXO} probably

3077: missed much hot gas at larger radii, perhaps as much as in the galaxy stars. 

3078: In fact, isothermal models fit data poorly,

3079: arguing again for non-gravitational heating. 

3080: 

3081: \subsubsection{Evidence from the cores of galaxy clusters.}

3082: 

3083: The same x-ray analysis has been applied to galaxy clusters. \citet{Bautz03}

3084: derive mass profiles and the central density slopes of 5 clusters

3085: with \textit{CXO}; 4/5 are consistent with $\Lambda$-CDM

3086: predictions. Only the cluster core may be in virial equilibrium. However,

3087: in some well studied cases there is clear evidence from x-ray

3088: images for transient core heating by shocks from sub-cluster

3089: infall \citep[for example][]{Markevitch02}.

3090: Such structures invalidate the isothermal assumption and may undermine

3091: the assumption of hydrostatic equilibrium. A further complication

3092: is a `cooling flow' in the centre of rich clusters and Egs

3093: \citep{Fabian94} whereby the densest hot gas cools onto the giant

3094: cD galaxies at the centres of clusters. In a few cases, central

3095: gas is measured to have half the temperature of the cluster mean.

3096: On the other hand, accretion rates up to 1000 M$_{\odot}$~yr$^{-1}$

3097: are predicted yet are not seen in soft x-rays. Radio jets from the

3098: AGN of the central galaxy appear to be reheating much of the gas and

3099: quenching SF (section \ref{sub:Ubiquitous-AGN-and}). 

3100: 

3101: The deflection and focusing of distant background light into multiple images

3102: (strong gravitational lensing) maps the distribution of

3103: mass across the cores of rich clusters.

3104: The pattern of images, for example cluster arcs, has been inverted to constrain

3105: the core mass (hence $\Upsilon$, see section 5 of \citealt{Bartelmann01}).

3106: DM can be mapped in sparser clusters by its

3107: shear of background images (weak lensing) to give a lower

3108: limit on the cluster core mass.

3109: HST's ACS camera covered three

3110: times the area of its predecessor with better spatial sampling, and

3111: covered targets out to 1.5 Mpc radii without mosaicing (which complicates

3112: derivation of a reliable shear pattern) on more distant clusters \citep[for example][]{Lombardi05}.

3113: From such a map, \citet{Clowe06} find that the DM precedes by 100

3114: kpc a distinct bow shock of hot gas falling through the `bullet

3115: cluster' 1E0657-56. The collisionless DM continues unimpeded while

3116: the x-ray gas is shocked hence delayed by the collision. DM in more

3117: than a dozen clusters has been assessed by lensing and compared to virial and

3118: x-ray measures. For example, the various mass estimates of cluster

3119: Abell 2218 out to radius 0.4 Mpc are $M{}_{\rm{stars}}=7.2\pm4.3\times10^{12}$

3120: \Ms, $M_{\rm{Xray}}=1.7\pm0.5\times10^{13}$ \Ms, and $M_{\rm{cluster}}=3.7\pm0.1\times10^{14}$ \Ms,

3121: implying that $M_{\rm{dm}}=3.5\pm0.2\times10^{14}$ \Ms\  and $\Upsilon=300\pm60$

3122: solar \citep{Squires96}; there is up to 100 times more mass than visible in stars \citep{Zwicky37}.

3123: The lensing estimate for $M_{\rm{cluster}}$ agrees with the result

3124: from virial analysis of member motions within the potential and from

3125: x-ray measurements, although lensing gives the mass inside a cone

3126: (along our l.o.s.) and the others give it in a sphere. The discrepancy

3127: is largest when the mass filament that contains the cluster is projected

3128: along our l.o.s. \citep{Squires96}.

3129: 

3130: Using accurate distances and a least-action infall into the Virgo cluster,

3131: \citet{Tully04} obtain total

3132: mass $1.2\times10^{15}$ \Ms, and $\Upsilon=850$, seven times

3133: higher than that found by the Virial theorem for the core. This

3134: analysis is unfeasible for other rich clusters.

3135: 

3136: 

3137: \subsection{Interactions with the host cluster}

3138: 

3139: 

3140: \subsubsection{Evidence for cluster effect on SF.}

3141: 

3142: The morphology--density relation ---

3143: many more Eg/S0 galaxies than Sgs in rich clusters

3144: and \textit{vice versa} in the low-density field \citep{Dressler80} ---

3145: is striking.

3146: Does the larger fraction of Egs in clusters arise from their birth

3147: (nature) or from evolution induced by the environment (nurture)? A

3148: strong clue is the `Butcher-Oemler effect' \citep{Butcher78}:

3149: galaxy clusters at moderate redshift $(z\sim0.3)$ have a larger fraction

3150: of photometrically blue galaxies than do clusters at low redshift.

3151: This finding indicates substantial evolution

3152: in cluster galaxies over the past $\sim5$ Gyr. Later studies have

3153: shown that some blue galaxies have standard Sg emission-line spectra.

3154: Others have spectra that are rare at present: very strong Balmer

3155: absorption lines without emission. Such resemble the composite of

3156: a main-sequence A and a K giant stellar spectrum, so are termed `E+A'

3157: or `K+A' galaxies. By inference, their SF ended within the

3158: past Gyr, leaving strong Balmer line relics. Their denotation as

3159: `post-star burst galaxies' is controversial because

3160: it implies that SF `ended with a bang', whereas some researchers argue that

3161: SF simply declined suddenly, i.e., was truncated (see section 7.2.2). 

3162: Balmer lines in some K+A galaxies are so strong

3163: that they clearly resulted from a star burst. But only if you catch

3164: it immediately after its burst can you distinguish unambiguously a

3165: post-burst galaxy from one that is fading with truncated SF.

3166: 

3167: High-resolution images with HST and ground-based adaptive optics systems 

3168: clarify that

3169: many blue galaxies are normal Sgs, the spiral fraction in clusters

3170: at $z>0.3$ much exceeds that today \citep[for example][]{Dressler97}, 

3171: and there was a larger fraction

3172: of interacting galaxies in $z>0.3$ clusters than at present

3173: \citep[for example][]{Lavery88,vanDokkum99}. Another

3174: clue is that the blue galaxies at higher-z tended to lie outside the

3175: cluster core, whereas most galaxies within the core are red even at

3176: high $z$. HST shows that the Eg fraction

3177: remains constant at $\sim15\%$ from $z=0.5$ clusters to the

3178: present; rather, it is the ratio of S0 to spiral galaxies that changes strongly

3179: \citep[for example][]{Dressler97}.

3180: Also, Egs at fairly high-z already seem to have a well-defined FP.

3181: In summary, the favoured scenario has cluster Egs already concentrated

3182: toward the cluster centre at early epochs $(z\sim2)$, whereas the

3183: Sg population has evolved rapidly in the past $\sim5$ Gyr as they

3184: converted to S0 galaxies while falling into the cluster.

3185: 

3186: More clues emerge from nearby clusters. \citet{Caldwell93} and \citet{Caldwell97}

3187: find early-type galaxies in nearby clusters

3188: with K+A `post-star burst' spectra that are generally too weak

3189: to classify as true K+A galaxies as defined in higher-z clusters.

3190: However, they \textit{do} show the same pattern of enhanced Balmer absorption lines

3191: without emission. Also, these K+A galaxies in nearby clusters

3192: tend to be rather low luminosity, i.e., they are much smaller galaxies

3193: than those seen at higher z. From their major study of Coma cluster

3194: galaxies, \citet{Poggianti04} argue that evolution in K+A galaxies

3195: can be explained by a `downsizing' effect, namely that galaxies

3196: terminating SF more recently are less massive. A possible explanation

3197: for how downsizing might work is that more massive Sgs form stars

3198: more rapidly, so later infalling massive `Sgs' no longer have

3199: gas to make the K+A effect.

3200: 

3201: 

3202: \subsubsection{Conversion of galaxy types.}

3203: 

3204: Two routes from spiral to S0 galaxies are proposed: tidal interaction/merger,

3205: and interaction between the Sg ISM and the hot $(10^{8}\rm{\rm{K}})$,

3206: diffuse intracluster medium (ICM). Both remove the ISM by stripping

3207: or by inducing copious SF. Tidal interaction invokes both processes,

3208: i.e., tides can remove significant gas and stars in a roughly equal

3209: mass encounter,

3210: but the interaction also drains much of the angular momentum

3211: from the gas into the galaxy centres to trigger a star burst \citep{Mihos96}.

3212: Tides tend to be strongest in \textit{low} velocity encounters so are

3213: disfavoured in galaxy clusters with their very high velocity

3214: dispersion. Thus, it has usually been assumed that tides and merging

3215: instead make Egs from disc galaxies. Yet there is evidence from

3216: high-resolution imaging of many merging/interacting Sgs in $z>0.3$

3217: clusters \citep[for example][]{vanDokkum99}. 

3218: So, an alternate scenario of `galaxy harassment' \citep{Moore96}

3219: destabilizes cluster galaxies by the aggregate of many rapid encounters

3220: with other galaxies and/or with the tidal field of the cluster. `Wet'

3221: (gas-rich) and `dry' (-poor) mergers are distinguished by the

3222: time required to organize the colliders into a spheroid: a dry merger

3223: builds a discy bulge quickly by violent relaxation, whereas secular

3224: processes (section \ref{sec:Bulges-of-Spiral}) can randomize the gas discs

3225: in a wet merger into a boxy spheroid in $1-2$ Gyr. 

3226: 

3227: The alternative to tides is the `ram pressure stripping' hypothesis

3228: of \citet{Gunn72}, who showed that an Sg falling through the hot

3229: ICM has its ISM stripped when the ram pressure exceeds the local restoring

3230: gravity of the disc.  Increasingly detailed

3231: numerical models have simulated stripping

3232: \citep[for example][]{Abadi99,Schulz01,Vollmer06}. Ram stripping

3233: should be effective when the density of the ICM is $\gtrsim10^{-3}$

3234: cm$^{-3}$ and the velocity dispersion of galaxies is $\gtrsim10^{3}$

3235: \kms, typical of rich clusters. What is seen? First, single-dish

3236: studies of many Sgs in nearby clusters show that their global HI

3237: content is much depleted compared to field Sgs \citep{Gavazzi87,Giovanelli85,Solanes01}.

3238: Radio interferometer mapping is more time consuming, so fewer such studies have been made.

3239: But they do resolve the HI in cluster

3240: Sgs and sometimes show a highly asymmetrical gas distribution,

3241: as if the ISM on one side of the galaxy is being compacted by

3242: ram pressure. Sometimes the radio continuum emission trails from the

3243: compacted side, and one can see a rim of HII regions concentrated

3244: along the squeezed edge \citep[or even ex-planar, as in][]{Kenney99},

3245: as if SF is being induced along the leading shock where the galaxy

3246: ISM plows into the cluster ICM \citep[for example][]{Gavazzi95,Crowl05}.

3247: 

3248: However, in the sparser Pegasus I cluster where the ICM density

3249: and galaxy velocity dispersion are low, ram pressure

3250: stripping should \textit{not} be occurring. Yet, \citet{Levy06} find

3251: cases where HI is depleted globally and even concentrates on one

3252: side of the galaxy where stars are forming. Other clues exist, but

3253: gas stripping from Sgs remains obscure.

3254: 

3255: Advocating that S0 galaxies in rich clusters like Coma are

3256: Sgs with stripped ISM raises questions.

3257: Most problematic is that HI-deficient Sgs in clusters appear to

3258: have systematically earlier morphological types and higher bulge/disc luminosity ratios

3259: than those with normal HI content \citep[for example][]{Dressler86}.

3260: How then can S0 galaxies with their apparently larger bulges be stripped Sgs? Perhaps,

3261: as SF exhausts gas, discs fade to increase the bulge/disc flux ratio.

3262: Too, \citet{Koopman98} have shown that the

3263: apparent excess of early-type stripped Sgs in clusters is caused at least partially

3264: by classification errors in morphology, due to the reduced SF rates.

3265: In addition, \cite{Caldwell99}

3266: have found examples of `bulges' of early-type galaxies in

3267: nearby clusters that resolve into star-forming or post-starburst regions

3268: (and thus are not `bulges' at all) when viewed

3269: in high resolution HST images.

3270: 

3271: There is also the question of cluster S0 galaxy ages. If these are really

3272: stripped Sgs, they would be younger than cluster Egs.

3273: \citet{Kuntschner00} studied both S0 and elliptical galaxies

3274: in the nearby Fornax cluster, and concluded that the S0s

3275: are indeed younger. On the other hand, \citet{Jones00} claim no age

3276: difference between these two types in clusters at $z\sim0.3$. Finally,

3277: if S0s form in clusters by stripping Sgs, how do they form in sparse

3278: environments?

3279: 

3280: A related question is: do most of the blue galaxies clustered

3281: at $z\sim0.5$ `passivate' into the non-SF galaxies seen in

3282: nearby clusters through a final major burst or by quenching/`strangling' their

3283: SF as their ISM is stripped? Do data suggest that

3284: their SF is simply truncated, or is there a last burst? The answer

3285: is controversial. The Canadian Network for Observational Cosmology

3286: consortium find no evidence for enhanced SF in $z\sim0.5$

3287: clusters \citep{Balogh99}, so argue that all Sgs in clusters are `strangled'. Conversely,

3288: the MORPHS consortium \citep[for example][]{Dressler04}, find strong

3289: Balmer line absorption that cannot be reproduced by the sudden truncation

3290: of SF; instead a star burst is required. In nearby clusters, \citet{Caldwell99}

3291: find that some galaxies at much lower luminosity than the bright star-forming

3292: galaxies at higher-z appear to be undergoing, or are seen right after,

3293: a star burst. So, although these galaxies are much smaller than big

3294: galaxies at higher-z with `K+A' signatures, they do seem to

3295: have undergone a last burst of SF not truncation.

3296: 

3297: In summary, data show that galaxies

3298: have evolved in both colour and morphology within clusters

3299: since $z\sim0.5$. Tidal

3300: damage and ram pressure stripping drive this evolution,

3301: and are being modelled with growing sophistication. However, much

3302: remains to be done to develop a coherent picture of rapidly evolving

3303: galaxy properties within a cluster.

3304: 

3305: 

3306: \section{New observational facilities}

3307: 

3308: A basic problem is the strong central

3309: concentration of galaxy light, especially in early-types. Light

3310: scattered from this point across the field of view of a spectrograph

3311: dilutes absorption line strengths. Calibrating this effect is complicated

3312: on a modern telescope because its altitude-azimuth mount (adopted

3313: universally for rigidity) rotates the field of view during a target

3314: track. The diffraction pattern of the mechanical supports of the secondary

3315: mirror also rotates across the field, injecting time-variable structure

3316: from the halos of bright sources within and even outside the field.

3317: Dusty optics at well ventilated observing sites scatter light over

3318: large field angles. Scattered light has an especially insidious effect

3319: when photometric profiles are averaged azimuthally, as is often done when one

3320: assumes axisymmetry. Diffraction gratings, including modern volume-phase

3321: holographic gratings, also scatter light unintentionally as it is

3322: dispersed.

3323: 

3324: Our atmosphere is scattering noise in many ground-based

3325: instruments. Shortward of $\lambda2000$ nm, the incoming wavefront

3326: can be `de-wrinkled' with an adaptive optics system to concentrate light. The ongoing

3327: retrofit of telescopes with adaptive optics will serve to resolve more distant

3328: stellar populations, allowing the powerful CMD discriminators to overcome

3329: the degeneracies of spatially integrated spectra. A `laser guidestar'

3330: allows an adaptive optics system to operate over almost the entire sky. Pixel SNR scales

3331: as the square-root of the exposure time when the signal is Poisson-noise

3332: limited. By reducing the size of the sky patch, an adaptive optics system reduces

3333: noise twenty-fold or more. 

3334: Complementing imagery is an IFS to obtain

3335: spectra at the limit of the adaptive correction.

3336: Hopefully, instruments on the very largest telescope(s) can be

3337: optimized to study both faint, distant blobs in

3338: the early universe \textit{and} bright, nearby puzzles.

3339: It is encouraging that the European

3340: Southern Observatory is building the Multi-Unit Spectroscopic Explorer (MUSE)

3341: \footnote{\url{www.eso.org/instruments/muse}} instrument whose 

3342: 1 arc-min$^2$ field of view will be corrected by adaptive optics.

3343: Such instruments could, for example, explore the DM content of an

3344: Eg through the kinematics of its dynamically relaxed globular clusters.

3345: 

3346: Many galaxies are red objects whose dominant stellar population

3347: emits mostly in the near-IR. The sky spectrum over $\lambda\lambda750-1500$

3348: nm is covered with numerous and variable 

3349: OH rotation-vibration band emission. The night

3350: sky is dark between, but

3351: light scatter within a spectrograph reduces this contrast.

3352: Work is therefore underway to modify optical fibres to suppress OH

3353: bands prior to wavelength dispersal. No technological barriers

3354: have appeared, suggesting that 99\% of the bands will be suppressed

3355: for another $4-8$-fold gain in SNR. 

3356: 

3357: After 2013, the James Webb Space Telescope (JWST) will attain high sensitivity

3358: long-ward of $\lambda1000$ nm with background

3359: set by dust emission and scattering (Zodiacal light). JWST

3360: will probe dust-shrouded regions of SF in nearby galaxies including

3361: the MWg. Missing within 5 years will be UV spectroscopy to assess the

3362: full impact of massive stars, so a relatively inexpensive 1.7m aperture

3363: World Space Observatory has been discussed.

3364: JWST development has already commercialized large-format (4 megapixel), edge-buttable,

3365: Presently, a single such detector cannot obtain simultaneous

3366: J+H-band spectra at the $R=4000$ required to work in the dark between OH 

3367: emission lines; a detector mosaic must be used. If one could use

3368: OH suppressing fibres, the full near-IR spectrum could be projected at

3369: $R=1500$ onto a single detector and a mosaic could be used to cover more 

3370: of the sky instead

3371: The ultimate such instrument in the next

3372: decade will have a laser adaptive optics system feeding deployable integral-field modules

3373: that are coupled by in-line OH sky-suppressing fibres to high-throughput

3374: spectrographs. 

3375: 

3376: For brevity, we mention only a few ground

3377: facilities under development for use by 2015.

3378: 

3379: The Large Synoptic Survey Telescope (\url{www.lsst.org}) is a 6.5m

3380: effective aperture telescope for rapid surveys (15 sec exposures)

3381: over 20,000 deg$^{2}$ and $\lambda\lambda320-1060$ nm in

3382: six bands. For example, it will map the DM content

3383: of galaxy clusters by weak lensing. Very cost-effective

3384: liquid mirror telescopes are

3385: also being developed for zenith strip surveys.

3386: 

3387: The 15-telescope Combined Array for Research in Millimeter-Wave Astronomy

3388: (CARMA, \url{www.mmarray.org}) in the N. hemisphere is currently,

3389: and the 64+ telescope Atacama Large Millimeter-Array (ALMA, \url{www.alma.nrao.edu})

3390: in the S. hemisphere will by 2012, provide $0.1-1$ arcsec resolution

3391: (to $0.01$ arcsec for ALMA) in millimetre and sub-millimetre

3392: CO emission, 5-50 pc resolution at 10 Mpc. GMCs and the dependence

3393: of the SF rate on environment can be studied throughout the local universe.

3394: 

3395: The Global Astrometric Interferometer for Astrophysics (GAIA,

3396: \url{www.rssd.esa.int/GAIA}) is a 1-meter aperture space telescope optimized

3397: for MWg kinematical studies. It will measure parallax distances

3398: (up to 90 million stars with better than 5\% accuracy and 1 million with 

3399: 1\% accuracy) and proper

3400: motions ($10^{9}$ stars, 1\% of the MWg population) out to a distance

3401: of $\sim10$ kpc from the Sun for giants and 1 kpc for dwarfs. In

3402: addition, it will obtain $R\sim10^{4}$ spectroscopy for radial velocities

3403: and limited metallicity information near $\lambda870$ nm for $\sim10^{8}$

3404: stars with $V<17.$ GAIA, combined with extensive ground-based followups,

3405: will revolutionize our view of the MWg.

3406: 

3407: This impressive technology will allow us to probe deeper into

3408: our nearest extragalactic neighbours, to understand how nature and nurture 

3409: have combined to present them to us today.

3410: 

3411: \section{Conclusions}

3412: 

3413: The quest to study nearby galaxies across the electromagnetic spectrum has brought

3414: these galaxies into sharp panchromatic focus to highlight their energy flows.  

3415: Contemporary surveys are establishing both the build up of

3416: mass/structure and the SF and chemical enrichment histories of

3417: the universe.  However, while impressive progress has been made, this review

3418: shows how precious little  we know about chemo-dynamical structures in

3419: galaxies and even less about how they operate.  Making assumptions about

3420: unobserved phase-space dimensions is always perilous.

3421: 

3422: We have focused on two themes related to nearby galaxies: 

3423: First, through the behaviour of visible tracers, what can we say about

3424: the distribution on their (presumably) dominant DM component?

3425: Second, what can

3426: we deduce about the internal structure, formation and evolution of their

3427: baryonic component?

3428: 

3429: Despite decades of study, the DM distribution within galaxies remains

3430: unclear. As we discussed, SB profiles in Egs are so steep that luminous

3431: tracers are too faint for accurate kinematic measurements at the

3432: radii where DM is presumed to dominate. Hence, except in the case

3433: of Local Group dwarf spheroidal galaxies, the case for DM in 

3434: individual Egs is still weak.

3435: and its distribution is barely constrained. In disc galaxies,

3436: DM is better established because RCs are often greatly extended in

3437: cold H I.  But there remains enough ambiguity in population 

3438: synthesis models of the visible matter (including uncertainty in 

3439: the low mass end of the stellar IMF) that the relationship between 

3440: mass and light is still much debated. Until these issues are resolved,

3441: details of the DM will depend on the uncertain visible mass 

3442: distribution. 

3443: 

3444: In regards to the internal chemo-dynamical structure and evolution of nearby

3445: galaxies, we have seen that despite decades of study, how

3446: elliptical and spiral galaxies are inter-related is still largely

3447: unknown.  The most basic questions about the 3-D

3448: structure of Egs remain unanswered, beyond the fact that

3449: massive ones are primarily flattened by anisotropic velocity

3450: dispersion while less massive ones tend to be flattened by

3451: rotation.  After much controversy over, and gradual refinements in, population

3452: synthesis models, we can now extract reliable \textit{light-weighted} mean

3453: ages and chemical abundances (of Fe, Mg and perhaps Ca).  We are only just

3454: beginning to progress to the far more difficult extractions of

3455: real SF and chemical enrichment histories of Egs, because

3456: these require accurate models of multiple spectral features at many wavelengths.

3457: 

3458: Our understanding of spiral galaxy bulges is even less certain

3459: than Egs, largely because study of the bulge requires identifying

3460: and removing the (thin and thick) disc component.  Information about bulge

3461: ages still comes largely from photometric colours, not spectroscopic

3462: line indices.  Recent understanding that pseudo bulges in late type spirals can constrain

3463: the hierarchical galaxy formation picture will be capitalized once it is determined 

3464: how often pseudo bulges arise from purely internal secular evolution and

3465: from minor mergers.

3466: 

3467: The core conclusion of this review is that, if we are to

3468: explore galaxy evolution scenarios such as $\Lambda$-CDM hierarchical

3469: merging by establishing galaxy properties out to high redshift,

3470: then we must improve considerably our understanding of nearby, present

3471: epoch galaxies. These systems benchmark high-redshift evolutionary

3472: studies.  The current generation of large ground-based

3473: telescopes and space-based facilities contribute by detailing individual

3474: galaxies.  Surface brightnesses of interest

3475: extend deep into the sky-dominated regime, unfortunately yielding

3476: only linear improvement in SNR with telescope diameter.

3477: Perhaps for this reason, 8-10+ meter telescopes are being

3478: used mostly to study high redshift galaxies at large lookback

3479: times.  However, of equal importance to increased aperture are improved detectors

3480: and more reliable control and removal of sky and galaxy light to

3481: enable study of galaxies into new regimes of low SB.

3482: Too, the development of integral field spectrometers has opened

3483: new possibilities for studying galaxy inner regions.  In short,

3484: there is much to be gained from glimpsing faint wisps within nearby ghosts.

3485: 

3486: \ack

3487: GC thanks the Reynolds Foundation for sabbatical support and

3488: Director Matthew Colless for hospitality at the Anglo-Australian Observatory

3489: during the start and finish of this review.

3490: We thank the referee for detailed and constructive comments.

3491: 

3492: \appendix

3493: \section{\label{sec:What-(Astro-)Physical-Processes}Relevant Astrophysical

3494: Processes}

3495: \setcounter{section}{1}

3496: 

3497: Understanding galaxy evolution requires familiarity with two branches

3498: of astrophysics: 1) the structure and evolution of stellar tracers

3499: of the dynamical and evolutionary state of galaxies, and 2) dynamics

3500: of self-gravitating systems. We first review basic stellar

3501: evolution theory. We discuss how to simulate the composite light of

3502: a coeval population, i.e. a cluster whose stars had the same birth.

3503: We introduce methods to compare the integrated spectra of model

3504: star clusters of unique age and chemical composition to the observed

3505: spectra, aiming to decipher the SF and chemical evolution histories of galaxies.

3506: Finally, we give a basic overview of self-gravitating systems, showing

3507: how their internal structure can be constrained by their

3508: observed light profiles and kinematics. Table \ref{tab:acronyms}

3509: lists the acronyms used in this paper.

3510: 

3511: \begin{table}

3512: \caption{\label{tab:acronyms}Acronyms used in this review.}

3513: \begin{indented}

3514: \item[]\begin{tabular}{ll}

3515: \br

3516: Acronym&Means\\

3517: \mr

3518: AGB&

3519: (stellar) asymptotic giant branch\tabularnewline

3520: AGN&

3521: active galactic nucleus\tabularnewline

3522: CBE&

3523: (stellar) collisionless Boltzmann equation\tabularnewline

3524: CDM&

3525: cold, dark matter\tabularnewline

3526: CMD&

3527: colour-magnitude diagram\tabularnewline

3528: CXO&

3529: Chandra x-ray Observatory\tabularnewline

3530: dIrr&

3531: dwarf irregular galaxy\tabularnewline

3532: dSph&

3533: dwarf spheroidal galaxy\tabularnewline

3534: DM&

3535: dark matter\tabularnewline

3536: Eg&

3537: elliptical galaxy\tabularnewline

3538: FP&

3539: (galaxy) fundamental plane\tabularnewline

3540: GMC&

3541: giant molecular cloud\tabularnewline

3542: HB&

3543: (stellar) horizontal branch\tabularnewline

3544: HI&

3545: neutral hydrogen\tabularnewline

3546: HII&

3547: ionized hydrogen\tabularnewline

3548: HST&

3549: Hubble Space Telescope\tabularnewline

3550: ICM&

3551: inter-cluster medium\tabularnewline

3552: IFS&

3553: integral-field spectrometer\tabularnewline

3554: IMF&

3555: (stellar) initial mass function\tabularnewline

3556: ISM&

3557: inter-stellar medium\tabularnewline

3558: IR&

3559: infra-red\tabularnewline

3560: LF&

3561: luminosity function\tabularnewline

3562: LSBg&

3563: low surface-brightness galaxy\tabularnewline

3564: MOND&

3565: Modified Newtonian Dynamics\tabularnewline

3566: Mpc&

3567: million parsecs\tabularnewline

3568: MS&

3569: (stellar) main sequence\tabularnewline

3570: MSTO&

3571: (stellar) main sequence turn-off\tabularnewline

3572: MWg&

3573: Milky Way galaxy\tabularnewline

3574: NSAR&

3575: non-stellar (chemical) abundance ratios\tabularnewline

3576: PA&

3577: position angle (on the sky)\tabularnewline

3578: RC&

3579: (galaxy) rotation curve\tabularnewline

3580: RGB&

3581: (stellar) red-giant branch\tabularnewline

3582: RGC&

3583: (stellar) red-giant clump\tabularnewline

3584: SB&

3585: surface brightness (of unresolved stars)\tabularnewline

3586: SDSS&

3587: Sloan Digital Sky Survey\tabularnewline

3588: SF&

3589: star formation\tabularnewline

3590: Sg&

3591: spiral galaxy\tabularnewline

3592: SNe&

3593: supernovae\tabularnewline

3594: SNR&

3595: signal-to-noise ratio\tabularnewline

3596: TFR&

3597: Tully-Fischer relation\tabularnewline

3598: URC&

3599: (galaxy) universal rotation curve\tabularnewline

3600: \br

3601: \end{tabular}

3602: \end{indented}

3603: \end{table}

3604: 

3605: 

3606: \subsection{\label{sub:Stellar-Evolution}Stellar Evolution}

3607: 

3608: 

3609: \subsubsection{Basic equations of structure.}

3610: 

3611: These have been established for

3612: decades, and are discussed in the elegant monograph by \citet{Schwarzschild58}.

3613: Throughout the star they characterize local hydrostatic equilibrium,

3614: mass conservation, energy production/conservation, and energy transport

3615: (either by convection, or radiatively from a temperature gradient).

3616: When coupled with an equation of state and after specifying

3617: boundary conditions (at a `surface' in wavelength averaged optical depth because

3618: the star has no sharp edge), one can integrate numerically the coupled

3619: differential equations for internal opacity and energy generation

3620: as functions of density, temperature, and chemical composition to

3621: obtain a self-consistent model.

3622: 

3623: 

3624: \subsubsection{Time dependence.}

3625: 

3626: No evolution equation depends explicitly on time. A star

3627: evolves because its chemical composition alters gradually through

3628: pressure confined fusions. While the ignition condition for each fusion

3629: is reached only deep in the star, convection can dredge

3630: up fusion products to alter the composition of the inert `envelope'.

3631: The main driver for evolution is that, as lighter elements

3632: fuse into heavier ones in the core, or later in a hot surrounding shell,

3633: the mean molecular weight of the gas mix declines

3634: to lower pressure. To maintain hydrostatic equilibrium, the core

3635: must contract to higher density and temperature. Because fusion rates

3636: are extremely sensitive to temperature, core energy release increases

3637: thereby expanding and cooling the envelope to brighten and cool the

3638: photospheric emission. Chemical composition usually changes very slowly

3639: compared to the dynamical timescale. So, the star's evolution can

3640: be traced numerically in time steps, each in hydrostatic equilibrium,

3641: with composition altered by the fusion that occurred during the previous

3642: step.

3643: On a modern computer it is trivial to evolve a spherical, non-rotating,

3644: unmagnetized star. While evolution models have become more sophisticated,

3645: \citet{Iben74,Iben91} remain excellent introductions.

3646: 

3647: 

3648: \subsubsection{The HR diagram.}

3649: 

3650: Before discussing how to derive galaxy ages from precise photometry

3651: of their individual stars and/or colours and spectra of their integrated

3652: starlight, a brief overview of stellar evolution is useful. Stellar

3653: evolution is characterized by the Hertzsprung-Russell diagram,

3654: wherein the star's effective temperature $T{}_{\rm{eff}}$, the

3655: temperature of the blackbody that radiates the star's photospheric flux per area

3656: from radius $R_{*}$,

3657: is plotted against the bolometric (full spectral) luminosity\begin{equation}

3658: L_{\rm{bol}}=4\pi R_{*}^{2}\sigma T_{\rm{eff}}^{4}\end{equation}

3659: Numerical models evolve a star in this diagram. As hydrogen fusion

3660: diminishes and the mean molecular weight of the core increases, the

3661: star evolves first to higher $L_{\rm{bol}}$ and lower $T_{\rm{eff}}^{4}$

3662: in the diagram. In practice it is impossible to measure either of

3663: these, because both require absolute spectrophotometry from the UV

3664: to the mid-IR. Instead, theory and observation intersect in the so-called

3665: colour-magnitude diagram (CMD), wherein the logarithm of a visible-band

3666: (or near-IR) flux is plotted vs. a visible-band (or near-IR) colour. Thus

3667: one needs accurate transformations between $L_{\rm{bol}}$ and the

3668: particular flux, and between $T_{\rm{eff}}^{4}$ and the particular

3669: colour, almost invariably by building a model of the stellar atmosphere

3670: to detail its photospheric spectrum.

3671: 

3672: Chemical composition, aside from providing input to determine

3673: a star's age, supplies an independent chronometer for galaxy evolution.

3674: Primordial gas from the Big Bang contains only hydrogen (mass fraction

3675: X), helium (mass fraction Y), and traces of other light elements (Li,

3676: Be, and B). Hence, details of the `heavy' element content (mass

3677: fraction Z) of a star's photosphere gives the degree of prior SF,

3678: subsequent evolution, and consequent heavy element production and

3679: expulsion (in winds and nova/SNe) prior to the star's birth in the

3680: enriched ISM.

3681: 

3682: Once photospheres of a sequence of stars have been so detailed, correlations

3683: with heavy element content may be obtained for grosser spectral features

3684: accessible to lower resolution spectroscopy or even filters selected

3685: carefully to measure photometric colours. Such data demand less telescope

3686: time, enabling study of fainter and more numerous targets in more

3687: diverse environments. As datasets are built, the original correlations

3688: are extended to less luminous stellar systems, to differing

3689: abundances, and to galaxies that have led different lives.

3690: 

3691: 

3692: \subsubsection{Colour-magnitude diagrams.}

3693: 

3694: Stellar evolution is complex as fusion shifts from core to shell hydrogen

3695: burning and then to helium core and eventually shell burning. As the

3696: dichotomy between an increasingly hot, dense core and an increasingly

3697: extended, cool envelope widens, equations of stellar structure become

3698: less certain. We give only the briefest overview of the important

3699: phases of stellar evolution. We aim to 1) acquaint the reader with

3700: how the various evolutionary phases can date a coeval population of

3701: stars, which is crucial to determining SF histories of galaxies,

3702: and 2) indicate which aspects of this program

3703: are robust and well-understood, and which aspects are uncertain and

3704: lead to an uncertain derived age. The underlying question is how accurately

3705: can CMDs of Local Group galaxies and the integrated spectrum of starlight

3706: from more distant galaxies constrain their SF histories?

3707: 

3708: \begin{figure}

3709: \begin{centering}\includegraphics[scale=1.8]{figure16.pdf}\par\end{centering}

3710: \caption{\label{fig:Theoretical-distribution-of}(a) Isochrones from

3711: the Geneva Observatory stellar evolution group

3712: of 3 different stellar populations, all with identical Salpeter IMF, and plotted for

3713: visible light filter bandpasses.

3714: The vertical axis is related to log luminosity and the horizontal

3715: to colour (hence temperature) with red (cooler) stars at right. The

3716: grey isochrone is for a young (1 Gyr old) stellar population of solar

3717: chemical composition. Black isochrones are for a 10 Gyr old population

3718: of solar composition (thicker) and of heavy element abundance 20\%

3719: of solar (thinner). (b) StarPop illustrates age-metallicity degeneracy.

3720: Two stellar populations arising from a 1 Gyr-long episode of constant

3721: SF are overlaid. Black plots a population of solar metallicity that

3722: ceased SF 1 Gyr ago. Gray plots a population of roughly half solar

3723: metallicity $(Z=0.008)$ that ceased SF 2 Gyr ago. RGB (right)

3724: and sub-giant branches largely overlap.}

3725: \end{figure}

3726: 

3727: The evolutionary track of a star --- the path over time in a CMD of a star of

3728: given mass and chemical composition --- can be computed.

3729: By combining tracks over

3730: a range of mass but with fixed chemical composition, one can predict

3731: the locus of all stars on a CMD at given age, an \textbf{isochrone}.

3732: An excellent Java Applet, StarPop

3733: \footnote{\url{http://astro.u-strasbg.fr/~koppen/starpop/StarPop.html}},

3734: synthesizes the evolution of a stellar population by sampling isochrones

3735: from a specified SF history. Figure \ref{fig:Theoretical-distribution-of}a

3736: illustrates the effects of age and chemical composition on stellar sequences in a CMD.

3737: 

3738: 

3739: 

3740: \subsubsection{Phases of stellar evolution.}

3741: 

3742: In Figure \ref{fig:Theoretical-distribution-of}a, note the long diagonal

3743: swath of young, solar chemical composition stars of different mass. Such `main

3744: sequence' (MS) stars are fusing hydrogen in cores. In contrast,

3745: the older solar-chemical composition population has a short MS because core

3746: hydrogen in the most massive stars has exhausted; only

3747: lower mass stars are still core fusing hydrogen.

3748: 

3749: Note the similar evolutionary sequences beyond the MS in both populations.

3750: Specifically, atop the MS is the MS turnoff (MSTO), a sequence of

3751: stars of increasingly higher mass and luminosity because the higher

3752: mass stars are increasingly evolved by depletion of their core hydrogen.

3753: In the CMD of a star cluster, the MSTO is our most effective chronometer.

3754: Because more massive stars have higher central temperatures and pressures,

3755: fusion exhausts fuel more rapidly. Hence, the mass

3756: of a star of known chemical composition that is just exhausting core

3757: hydrogen can yield a stellar age reliable to $\sim1$ Gyr on populations

3758: older than $1-2$ Gyr. 

3759: 

3760: Subsequent, shorter phases of stellar evolution are:

3761: \begin{itemize}

3762: \item The red giant branch (RGB), again a sequence of increasing mass, now

3763: increasingly evolved. It illustrates the increasing dichotomy between

3764: hot core and cool low-density envelope as hydrogen fusion switches

3765: to a shell. RGB stars are crude chronometers for populations older

3766: than 1 Gyr because stars over a wide mass range funnel into the same

3767: part of the CMD. The most important spectral feature for RGB

3768: stars is the triplet of singly ionized calcium at $\sim\lambda800$

3769: nm; it is useful even at lower spectral

3770: $R\sim6000$ that washes out Fe lines. \citet{Gallart05} review the

3771: uncertainties and conclude that, for a resolved population, the integrated

3772: SF rate from its start until $\sim2$ Gyr ago can be estimated

3773: to within a factor of four from counts of RGB and asymptotic giant

3774: branch (AGB, see below)

3775: stars, provided that temporarily over-luminous AGB long-period variables

3776: are accounted for. 

3777: \item Subgiant-branch stars, evolving relatively slowly toward the

3778: RGB by burning H in a shell around an inert He core. All stars with

3779: MS life longer than 2 Gyr experience this stage.

3780: \item The red giant clump (RGC) is the helium core-fusion MS

3781: for stars 1-10 Gyr old. Helium ignites suddenly because some electrons

3782: in the core became quantum statistically degenerate when

3783: the star ascended the RGB, augmenting the normal gas-pressure adjustments

3784: for hydrostatic equilibrium

3785: with a temperature-insensitive, incompressible `lattice' of

3786: electrons. The result is a discontinuous loss of envelope mass and

3787: a readjustment of the star's interior structure. As it ascends the

3788: RGB, the star loses mass from its outer envelope at a rate that is

3789: ill defined and intractable theoretically, rendering all evolutionary

3790: calculations beyond the early part of the RGB uncertain.

3791: \item Horizontal branch (HB) stars also burn He in their cores, are low

3792: mass, older than 10 Gyr, metal-poor, and have a more extended range of

3793: temperatures than RGC stars.

3794: RR Lyrae stars are examples, recognized individually because of their

3795: variability. The ratio of red to blue HB stars constrains the metallicity

3796: and age. \citet{Gallart05} review the potential to date ages with

3797: the RGC and HB, and conclude that they work reasonably well for $1-3$

3798: Gyr ages. 

3799: \item The asymptotic giant branch (AGB). After core helium burning at the

3800: RGC, the star burns both helium and hydrogen in separate shells with helium fusion

3801: unstable.

3802: AGB stars are the most luminous known. They are chronometers

3803: for populations older than 1 Gyr, but are less accurate when younger

3804: because of uncertain wind-driven mass loss, thermal

3805: pulses from helium shell fusion flashes, changes in photospheric abundances

3806: from heavy element dredge-up (in particular, carbon), and mixing.

3807: Uncertain contribution of AGB stars can alter the fluctuations

3808: of the SB of galaxy starlight. If the RGB tip is reached below $\sim$8 \Ms,

3809: the stellar envelope is gone and nuclear burning soon ceases.

3810: \item A short, bright `planetary' nebula phase wherein the ejected

3811: envelope of a $0.8-3$ \Ms\ star is photoionized by UV from

3812: the exposed helium, carbon or oxygen core. Thereafter, it fades as a low luminosity

3813: white dwarf whose

3814: core is supported by a barely cooling degenerate electron lattice.

3815: White dwarfs are too faint to study beyond the MWg.

3816: \end{itemize}

3817: The black sequences in Figure \ref{fig:Theoretical-distribution-of}a

3818: show the importance of chemical composition in stellar evolution.

3819: Both interior (opacities, hence radiative energy transfer, and mean molecular weight, hence

3820: internal pressure) atmospheric

3821: (spectral line strengths, hence the emergent radiation field) properties

3822: depend on chemical abundances. As a result, the metal-poor (thinner line)

3823: MS lies blueward (hotter) of the MS of the metal-rich isochrone (thicker

3824: line). The message is that an accurate CMD (from photometric imagery),

3825: an independent measure of its chemical composition (from high resolution

3826: spectra of the absorption lines in individual stars), and a reliable

3827: theoretical isochrone, can be combined to date the population. But,

3828: isochrones are only as accurate as stellar evolution, which is uncertain

3829: in several areas. There are uncertainties in

3830: \begin{itemize}

3831: \item key nuclear reaction rates $^{12}C(\alpha,\gamma)^{16}O$,

3832: $^{14}N(p,\gamma)^{15}O$, and even the triple-$\alpha$ reactions

3833: are either poorly constrained or have been revised recently to alter

3834: evolution timescales \citep{Weiss05}. The problem for

3835: determining astrophysical cross-sections is the difficulty in producing

3836: sufficiently low-energy but high-luminosity accelerator beams.

3837: \item opacities that determine radiative energy transport.

3838: Widely used opacities are from the Opacity Project

3839: \citep{Seaton94,Seaton96} and the OPAL \citep{Iglesias96} consortium.

3840: A larger opacity lowers the H fusion rate and core temperature. 

3841: \item the chemical composition of even the Sun, the basis of abundance analyses

3842: of all other stars. Recent redetermination by \citet{Asplund05a}

3843: indicates that the abundance of O and other major elements must be

3844: reduced $\sim$1.5-fold, a controversial result because it worsens

3845: agreement between theory and helioseismology spectra of the solar

3846: interior \citep[for example][]{Bahcall05,Antia06}. Increasing helium

3847: abundance Y increases $T_{\rm{eff }}$ and $L_{\rm{bol}}$ and

3848: decreases the width of the MS.

3849: \item convection, which does not seem to have even a mid-term solution.

3850: We know so little about it that models parametrize it by a

3851: convective scale height, the typical distance travelled by a convective

3852: cell in units of the local pressure scale height. Because convection

3853: transports energy in the outer layers of the Sun, its scale height sets the solar radius.

3854: So, this scale is fixed \textit{ab

3855: initio} in a 4.6 Gyr model to reproduce the solar radius, and its

3856: value is assumed to apply universally to stars whatever their

3857: internal structure, increasing the core mass. Our ignorance

3858: is further compounded by convective overshoot: the scale is increased

3859: to account for the inertia of bubbles that turn around not where their

3860: acceleration reaches zero but beyond, at zero velocity. Overshoot

3861: is clearly highly uncertain, but appears to be necessary to match

3862: model isochrones with the CMDs of intermediate-age star clusters \citep[for example][]{Kozhurina97}.

3863: Finally, convection plays a huge r\^ole by intermixing layers that

3864: are otherwise stratified into different chemical composition. For

3865: example, deep mixing transports fresh hydrogen into the core, prolonging

3866: a star's life. Models for intermediate-mass stars are compared by

3867: \citet{Dominguez99}, who find 10\% variations in H-burning lifetimes

3868: and 30\% for He core-burning.

3869: \item handling mass loss. As a star ascends the RGB, its envelope

3870: distends and becomes less bound. An AGB star loses mass through either

3871: a gradual wind or sudden thermal pulses. The empirical Reimer's Law \citep{Schroder05}

3872: estimates gradual mass loss by a wind from an 

3873: RGB star, but data are hardly reliable enough to predict the star's

3874: mass when it ignites helium in its core. Thus, all evolutionary

3875: calculations become more uncertain once the star ascends the RGB.

3876: Specifically, the location of a star in the HR diagram during its

3877: core helium burning, HB phase depends on the envelope mass (whereas

3878: the mass of the helium core, hence the star's luminosity, is relatively

3879: insensitive to initial mass): stars with lighter envelopes have hotter

3880: $T_{\rm{eff}}$, so are bluer. Uncertain mass loss confounds prediction

3881: of a star's position along the HB. Hence, isochrones that include

3882: HB and post-HB phases have made basic assumptions about mass loss

3883: that usually force the morphology of the HB to agree with that

3884: observed in star clusters.

3885: \item element diffusion from gravitational settling in the outer, radiative

3886: layers that can mask the star's true chemical composition.

3887: Helium diffusion in the core will alter the star's lifetime. An independent

3888: check on its efficiency comes from other elements, \citet{Straniero97}

3889: suggesting those not burning at H-burning temperatures, Fe or Ca,

3890: as candidates and both measurable in Local Group galaxies down to the MSTO

3891: with large telescopes.

3892: \item the unobserved stellar rotation that augments hydrostatic equilibrium

3893: and breaks the spherical symmetry of the star to greatly complicate models.

3894: Rotation is ignored when many models are generated for an isochrone.

3895: Rotation alters deep mixing hence chemical abundances.

3896: \item Coulomb effects that soften the equation of state. 

3897: They reduce MSTO ages by 10-15\%,

3898: and ages derived from B- and V-band colour differences by 40-50\%.

3899: \end{itemize}

3900: These uncertainties grow at more advanced evolutionary stages as the

3901: internal structure of the star becomes more extreme. Hence, the most

3902: robust estimate of the age of a star cluster remains the $T_{\rm{eff}}$

3903: and/or luminosity of the MSTO, as determined from a CMD. As we will

3904: see, when individual stars cannot be resolved, one can still infer

3905: MSTO characteristics by deciphering the integrated spectrum of galaxy

3906: starlight.

3907: 

3908: 

3909: \subsubsection{Sensitivity to basic parameters.}

3910: 

3911: Broad bandpasses bluer than V cannot find T$_{\rm{eff}}$ accurately

3912: because the spectrum is blanketed by a forest of absorption

3913: lines whose strengths vary with chemical abundances and surface gravity

3914: $g=GM/R^{2}$. Larger telescopes and more sensitive detectors have

3915: supplanted photometry with spectral determination of stellar

3916: parameters. For 5000 K$\le$T$_{\rm{eff}}$$\le$8000 K, Balmer line strengths yield

3917: accuracy $\pm50-80$ K. Surface gravity can

3918: be established to a multiplicative accuracy of 1.5--2.5 by its pressure

3919: modification of some line profiles (mainly Fe, Ca I,

3920: Mg Ib triplet at $\lambda$520 nm, and CN bands at $\lambda$380 +

3921: $\lambda$410 nm) \citep[for example][]{Fuhrmann97}. \citet{Lebreton01}

3922: discusses how star diameters derived from interferometric measurements

3923: have calibrated to internal accuracy $\sim1\%$ a combined empirical/model

3924: atmosphere Infrared Flux method to deliver $T_{\rm{eff}}$ for A-F

3925: dwarf and giant stars \citep{Blackwell98} and, an SB method with

3926: comparable accuracy \citep{DiBenedetto98}. Recent work 

3927: \citep[for example][]{Bessell98} improved

3928: treatments of atomic/molecular line blanketing in O-K and

3929: A-M stars.

3930: 

3931: 

3932: \subsubsection{Sensitivities to variations in Y and Z.}

3933: 

3934: \citet{Lebreton01} reviews measured chemical abundances; \citet{Wallerstein97}

3935: review the derived patterns. Abundance estimates are less accurate for hot

3936: and young stars because most of their ionization structure is in the

3937: UV, and because spectral features are blurred by rapid stellar rotation.

3938: Elements to study include

3939: \begin{itemize}

3940: \item O, Mg: products of SNe II that probe contributions from $>20$ \Ms\ 

3941: and trace chemical mixing in globular clusters.

3942: \item Si, Ca, Ti, Cr: also products of SNe II but less strongly weighted

3943: toward more massive progenitor stars. Combined with O and Mg, these

3944: `$\alpha-$elements' (even-proton number nuclei) probe the distribution

3945: of initial masses (the so-called initial mass function, IMF) \citep{McWilliam97}.

3946: \item Mn, Co: from SNe II, with yields depending on metallicity of

3947: the progenitor star.

3948: \item Eu: indicate rapid (r-process) neutron pickup (compared to the rate

3949: of $\beta$-decay) nucleosynthesis in SNe II.

3950: \item Y, Zr, Ba, La: indicate slow (s-process) neutron pickup nucleosynthesis

3951: in lower-mass AGB stars.

3952: \item Fe, Ni: track metallicity from SNe Ia. Fe also provides

3953: parameters such as $T_{\rm{eff}}$ and $g$.

3954: \end{itemize}

3955: Detailed abundances constrain the SF history. $\alpha$-elements come

3956: preferentially from hydrostatic burning in massive stars before their

3957: SNe II make $r$-process elements. Figure \ref{fig:SNyeildsim} \citep[Fig. 1 of][]{Bland-Hawthorn04}

3958: shows how successive generations of low-mass stars enrich by SNe II.

3959: Roughly 1 Gyr after the star burst, white dwarf deflagration/detonation

3960: SNe Ia (from mass

3961: accretion in a binary star system) make more Fe-peak elements. An

3962: observed overabundance of $\alpha$-elements compared to Fe-peak elements

3963: implies a brief or recent SF episode \citep{Trager00,Trager04}; their

3964: detailed abundances depend on the mass of the supernova progenitor. During

3965: this time, AGB stars make odd-proton nuclei in the $s-$process. The

3966: ratio of $\alpha$/Fe-peak element abundance decreases as a star burst

3967: progresses. So the relative abundances of $r-$ to $s-$process elements

3968: gives the relative importance of SNe Ia to AGB stars, hence the relative

3969: yields of higher and lower mass stars.

3970: 

3971: \begin{figure}

3972: \begin{centering}\includegraphics[scale=0.3]{figure17.pdf}\par\end{centering}

3973: \caption{\label{fig:SNyeildsim}

3974: \citep[From][used with permission of Publications of the Astronomical Society of Australia, 

3975: the authors and the publisher CSIRO.]{Bland-Hawthorn04}

3976: How successive generations of low-mass stars

3977: are enriched by just SNe II, as traced by europium production (r-process).

3978: The initial wide scatter in abundances at left converges to a universal

3979: value at right. The stars are formed with a Salpeter IMF and the element

3980: yields are from \citet{Tsujimoto98}. The upper bound comes from lower-mass

3981: SN, the lower from higher-mass. From left to right, \fullsquare\ indicates

3982: the number of prior enrichments: 1, 10, 100, 1000. Just 10

3983: enrichments from high-mass SNe suffice to enrich a cloud to $[Fe/H]=-2$.}

3984: \end{figure}

3985: 

3986: 

3987: 

3988: \subsubsection{Uncertain isochrones.}

3989: 

3990: 

3991: 

3992: The pace of stellar evolution is exponential with mass. To combine tracks

3993: to form an isochrone, one must interpolate in

3994: the $(\log M,\log X)$ plane (X is the core hydrogen mass abundance)

3995: by assuming that stellar evolution is continuous in this plane (it

3996: is not), and that tracks of stars of slightly different mass are

3997: always proximate in the CMD (also untrue in some mass ranges). Isochrones

3998: may be interpolated physically only when homology relations can

3999: scale stellar properties. Homology is broken by mass loss and

4000: when core processes switch

4001: drastically over a small mass range such as the dominance of CNO over

4002: proton-proton fusion. Interpolation to intermediate

4003: masses then yields invalid tracks through unphysical parts of the CMD.

4004: Homology relations anyway predict only indirectly observable interior

4005: --- not observable photospheric --- properties. Finally, all models

4006: ignore important evolutionary influences such as rotation, binary

4007: mass transfer, and magnetic fields.

4008: 

4009: Isochrones differ slightly depending on how RGB and HB convections

4010: are handled \citep{Gallart05}. For example Geneva

4011: (\url{http://obswww.unige.ch/~mowlavi/evol/stev\_database.html})

4012: and Padova (\url{http://pleiadi.pd.astro.it/}) group zero-age isochrones span observations

4013: of both halo and disc stars only if temperatures are set 130--250 K hotter than observed.

4014: 

4015: \subsubsection{Mass functions.}

4016: 

4017: Stellar $M/L\equiv\Upsilon$ is found empirically to vary from $4\times10^{-5}$

4018: (O5 star) to 34 (M5). This mass-luminosity relation is fundamental,

4019: yet its measurement is limited to those stars in suitable binary orbits

4020: that are also near enough to us to have a precisely measurable parallax;

4021: a 10\% parallax error produces a 5\% error in $T_{\rm{eff}}$. 

4022: 

4023: Models must specify a stellar birth-rate. The usual simplification

4024: forms stars in one burst, so have a single age, metallicity, and initial

4025: mass function (IMF) that together define a so-called simple stellar

4026: population for the entire region observed. The

4027: IMF is the probability of finding a star of given mass;

4028: it thus also sets the mass fraction of the population that

4029: is returned to the ISM by evolved stars. As a distribution, its use

4030: implies an unavoidable Poissonian sampling variance

4031: \citep[see][for a correction procedure]{Cervino02}; for example,

4032: the predicted colours of a $10^{5}$ \Ms\  cluster (which is $\sim6$

4033: times the mass of the most massive OB star association known in the

4034: MWg, Cygnus OB2) disperse by 3-10\% depending on the assumed IMF and

4035: on whether or not a few rare but luminous stars are included. 

4036: 

4037: IMF's are based on censuses of nearby SF regions. However, these regions

4038: make predominantly low mass stars, are subject to variance, and are

4039: complicated observationally by unresolved binary stars. The \citet{Salpeter55}

4040: power-law IMF $\zeta(m)\propto m^{-\alpha}$ with $\alpha=1.7$ is

4041: standard and has been assumed in most studies. However, \citet{Kroupa01}

4042: summarizes evidence that this IMF predicts more low-mass stars than

4043: are observed in MWg and LMC open star clusters; he provides an alternative,

4044: multiple-part IMF from fits to those regions. The mean stellar mass

4045: with this IMF is $0.36$ \Ms; half of the mass is in stars $0.01\le m\le1$ \Ms\  

4046: and the other half is in $1\le m\le50$ \Ms. The variances

4047: arise mostly from unresolved binaries. 

4048: 

4049: 

4050: 

4051: \citet{Bruzual03} show that $\Upsilon$ derived from different

4052: plausible IMF's varies by multiplicative factor $\sim1.5$ for

4053: up to 1 Gyr after a star burst, increasing to multiplicative factor

4054: $\sim2.5$ for ages up to 10 Gyr. Another uncertainty on the low

4055: end of the IMF is that the current census with mass $<0.5$ \Ms\ 

4056: is incomplete beyond 5 pc \citep{Henry97}. The high end of the IMF

4057: (especially for Wolf-Rayet stars) has uncertain normalization and

4058: highest variance from observations \citep[see for example Fig. 8 of][]{Bruzual03};

4059: this is irrelevant for models of early-type galaxies that have not

4060: formed many stars recently. In their high-SNR study of 25 such galaxies

4061: (see section \ref{sub:Global-scaling-parameters}), \citet{Cappellari06}

4062: compare the dynamically determined $\Upsilon$ with the values implied

4063: from population synthesis \citep{Vazdekis96} to fit the observed

4064: absorption strengths. They obtain better agreement between the two

4065: methods across their sample with the \citet{Kroupa01} IMF; its fewer

4066: low-mass stars compared to Salpeter reduces the derived $\Upsilon$

4067: and hastens the subsequent evolution of the star cluster.

4068: 

4069: 

4070: \subsubsection{Building a model galaxy.}

4071: 

4072: Having considered the evolution of a single star (evolutionary track

4073: in the HR diagram) and of a stellar system (isochrone), we can

4074: now model a simple stellar population, i.e., a stellar system

4075: that is co-eval with unique chemical composition. A grid of SSP's,

4076: ranging in age and chemical composition, forms the basis for comparison

4077: with real galaxies. To create the spectrum of an SSP, one selects

4078: an isochrone of desired age and metallicity and an appropriate IMF.

4079: Next, at each point in the HR diagram, the number of stars (weighted

4080: according to the isochrone and the IMF) is calculated, and a stellar

4081: spectrum is selected with the appropriate $T_{\rm{eff}}$, $\log g$, and metallicity. Finally,

4082: the spectrum is weighted by the number of stars at that point and

4083: by luminosity, and combined with those at all other points in the

4084: diagram to synthesize the spectrum of the particular SSP. \citet{Bruzual03}

4085: detail this procedure. Naturally, its success depends on having a

4086: spectral database that covers a wide range in $T_{\rm{eff}}$, $\log g$,

4087: and chemical composition. Because of the difficulty of modelling the

4088: emergent spectra of stars with millions of atomic and molecular transitions,

4089: it has been traditional to use `libraries' of uniform-quality

4090: spectra of many stars whose $T_{\rm{eff}}$, $\log g$, and metallicity

4091: has been determined from a high dispersion abundance analysis. Recent

4092: libraries, each comprising spectra of $\sim$1000 stars and ranging

4093: widely over atmospheric parameters, are

4094: 

4095: \begin{itemize}

4096: \item The Indo-US library of coude-feed stellar spectra \citep{Valdes04}

4097: at \url{www.noao.edu/cflib}, covers $\lambda\lambda$346 - 946 nm

4098: at resolution $\sim$$\lambda$0.1 nm FWHM for 885 stars.

4099: \item ELODIE \citep{Moultaka04} spans $\lambda\lambda$400 - 680 nm at

4100: high dispersion. Based on stars with well-determined atmospheric

4101: parameters, version \url{http://atlas.obs-hp.fr/elodie} has $>10,000$

4102: spectra.

4103: \item MILES \citep{Sanchez06} contains spectra at resolution

4104: $\lambda$0.23 nm FWHM and covers $\lambda\lambda$352 - 750 nm. 

4105: \end{itemize}

4106: Empirical libraries have limits. It is difficult

4107: to find stars that range over sufficient $T_{\rm{eff}}$, $\log g$,

4108: and metallicity whose atmospheric parameters are known from fundamental

4109: high-dispersion spectral analysis. This is particularly true for hot

4110: stars, because only metal-rich examples exist in the Solar Neighbourhood and

4111: it is now clear that the mean

4112: abundance ratios of stars in Egs are non-solar (see section 3.6).

4113: Establishing atmospheric parameters for

4114: stars in spectral libraries is also problematic, because the high dispersion analyses

4115: were made by different investigators using different model atmospheres

4116: and $T_{\rm{eff}}$ scales from different photometric colours.

4117: 

4118: Models of stellar atmospheres have increased in sophistication through

4119: great increases in both computational power and the number of included

4120: atomic and molecular transitions. Hence, using a library of synthetic

4121: stellar spectra is increasingly attractive. Theoretical stellar spectra

4122: can be calculated for any chemical prescription. For full consistency,

4123: one should first model an atmosphere based on the assumed chemical

4124: composition, and then calculate the detailed emergent spectrum. While

4125: calculating the model is very time consuming, use of theoretical spectra

4126: is now widespread especially for constructing the hot stars in young

4127: stellar populations. An example is the library of 1654 theoretical

4128: spectra, sampled at $\lambda$0.03 nm and covering $\lambda\lambda$300

4129: - 700 nm, that is described in \citet{Martins05}\footnote{\url{www.astro.iag.usp.br/~lucimara/library.htm}}.

4130: 

4131: Two approaches compare composite spectra of SSP's, gridded by age

4132: and metallicity, to the observed spectrum to determine which

4133: combination of SSP's matches best.

4134: 

4135: \begin{enumerate}

4136: \item Focus on absorption lines in the model grids with particular sensitivity

4137: to age and/or metallicity, then find their best match. The preeminent

4138: example is the Lick system of spectral indices \citep[for example][]{Faber85}.

4139: This approach has several attractions:

4140: 

4141: \begin{itemize}

4142: \item It is transparent to which specific feature(s) determine

4143: age and metallicity.

4144: \item It is simple to compute a two-index diagram that separates age and

4145: metallicity effects (for example H$\beta$ versus the Fe5270 index),

4146: plot the grids that connect the various SSP models, overplot galaxy

4147: measurements, and then interpolate ages and metallicities for a large

4148: galaxy sample.

4149: \item With model atmospheres, one can probe precisely which transitions

4150: of various elements form the index, either through the central passband

4151: or through the two continuum bands that straddle it in wavelength.

4152: \end{itemize}

4153: \item Use $\chi^{2}$ minimization to fit the entire spectrum to models.

4154: This approach is problematic because a typical spectrum contains more than 1000

4155: independent pixels, so computation is intensive and impractical for

4156: the numerous $(\sim10^{6})$ spectra from an extensive survey such

4157: as the SDSS. Moreover, spectral information is highly redundant, with

4158: many features duplicating the overall degenerate sensitivity to age

4159: and metallicity. If the SNR is low, co-adding many redundant features

4160: is advantageous, although populations would then not likely be isolated.

4161: If the SNR is high, information in the spectrum can usually be reduced

4162: to a few parameters that are chosen judiciously. Various strategies

4163: are possible:

4164: 

4165: \begin{itemize}

4166: \item Pre-select line strength indices, as in 1) above. 

4167: \item Use principal component analysis (\citealt[for example][]{Madgwick03,Ferras06})

4168: to find the minimum eigenvector set that contains most of the spectral

4169: information. Solving for the eigenvalues that characterize a given

4170: observed spectrum yields enormous compression. However, spectral

4171: evolution of stellar populations is very non-linear in time. Thus,

4172: for a set of emission-free spectra of different ages (to avoid the

4173: vexing emission-line `contamination' of absorption features),

4174: the galaxy spectrum is synthesized by combining a very

4175: young with a very old population that may be irrelevant to that galaxy's

4176: SF history.

4177: \item Compress spectra through a set of weighting vectors to match a set of realistic

4178: population parameters (for example a dozen ages spaced logarithmically, each with

4179: a metallicity). \citet{Panter03} have used

4180: such compression (acronym MOPED) to extract SF histories from huge

4181: numbers of SDSS spectra. Of concern is that this approach fits data

4182: with a highly compressed $\chi^{2}$ minimization that may fail to

4183: duplicate faithfully the observed spectrum. While the same criticism

4184: can be levelled at the spectral index approach, mismatches between

4185: models and data are there more obvious. For example, plotting the

4186: age-sensitive Lick H$\beta$ index against the Fe-sensitive Fe5270

4187: index on a model grid with various ages and metallicities but using

4188: solar abundance ratios, yields a single metallicity. For massive Egs,

4189: plotting H$\beta$ against Lick Mg b indices (Mg b measures primarily

4190: a Mg-sensitive feature) yields a larger metallicity. Such results

4191: have indicated clear non-solar abundance ratios (NSAR) in Egs \citep[for example][]{Worthey04,Trager00,Peletier89},

4192: requiring consistent NSAR models \citep{Thomas03}. Such transparent

4193: disagreement between model and data led to important insights on the

4194: SF/chemical evolution histories of galaxies; it is far less clear with the MOPED and

4195: principal component techniques.

4196: \end{itemize}

4197: \end{enumerate}

4198: 

4199: \subsection{\label{sub:Gravitational-stellar-dynamics}Gravitational stellar

4200: dynamics}

4201: 

4202: Goals of gravitational dynamics are to

4203: 

4204: \begin{enumerate}

4205: \item Understand the phase-space distribution of stars in galaxies. The

4206: observed 2D projected stellar motions and SB distribution constrain

4207: the 3D shape and motions in the potential. The stars, with the inferred

4208: DM, establish the gravitational potential of the galaxy. (Gas is generally

4209: negligible in gravity, simply acting as a tracer that may not be in

4210: equilibrium.)

4211: \item Establish the stability and secular evolution of the stellar and DM

4212: distributions in galaxies. How do these two interact over time in

4213: `isolated' galaxies? 

4214: \item Establish the consequences of gravitational interactions between galaxies. 

4215: \end{enumerate}

4216: One must recognize that a

4217: self-gravitating system differs greatly in physics from a

4218: charged plasma, perhaps making gravitational dynamics seem

4219: obscure and counter-intuitive. 

4220: 

4221: First, unlike in a plasma with equal numbers of $\pm$ charges, gravity

4222: only attracts; no equivalent to Debye shielding limits interactions

4223: to roughly the inter-particle separation. Thus, interactions in a

4224: gravitational system are very long-range; indeed, the sum of weak

4225: ones from very distant stars dominates over nearby

4226: two-body encounters. Surrounded by $\gtrsim10^{5}$ stars, each star

4227: sees a smooth potential and would need longer than the

4228: age of the Universe to exchange significant energy with others;

4229: self-gravitating systems are nearly collisionless. This implies that,

4230: if the potential has been stable since a galaxy formed, then a

4231: star retains memory of its original orbit and its observer can be

4232: a `galactic archaeologist'. Conversely, if evidence suggests

4233: that certain orbit families have altered substantially, then by implication

4234: the galactic potential was disrupted significantly at least once in

4235: the past to scatter stars into new orbits.

4236: 

4237: Second, the stellar mean free path length exceeds greatly the system diameter,

4238: so its equation of state, pressure, and temperature cannot be defined.

4239: In contrast, charged plasma particles remain in a local equilibrium

4240: characterized by the ideal gas law. 

4241: 

4242: As will soon be clear, solving the equations of a self-gravitating

4243: system is daunting.

4244: We touch only on key aspects, and

4245: refer the reader to the classic monograph of \citet{Binney87} to

4246: understand the full problem.

4247: We first introduce the distribution function $f(\mathbf{x},\mathbf{v},t)$,

4248: the fine-grained probability of locating stellar mass in six-dimensional

4249: phase-space. Without collisions, the mass density within

4250: this piece of phase space is invariant $df/dt=0.$ Hence, the coupled 

4251: Boltzmann and Poisson equations in gravitational potential $\Phi$

4252: \begin{eqnarray}

4253: \begin{array}{c}

4254: \frac{df}{dt}\equiv\frac{\partial f}{\partial t}+\mathbf{v}\bullet\boldsymbol{\bigtriangledown}f-\boldsymbol{\bigtriangledown}\Phi\bullet\frac{\partial f}{\partial\mathbf{v}}=0 \\

4255: \\

4256: \bigtriangledown^{2}\Phi(\textbf{x},t)=4\pi G\int f(\mathbf{x},\mathbf{v},t)d^{3}\textbf{v}

4257: \end{array}

4258: \end{eqnarray}\label{eqn:poisson}

4259: describe the dynamics

4260: of collisionless, self-gravitating galaxies.

4261: Seven independent variables make $f$ hard to solve.

4262: Instead, one often obtains insights from spatial

4263: or kinematical moments, yielding three partial differential equations.

4264: But, moments are \textit{global} averages over

4265: the \textit{entire} distribution, and there certainly \textit{is} considerable

4266: material (both DM and possibly unsettled baryons) in the faint

4267: outer parts of galaxies, with unknown kinematics.

4268: In fact, this material is very important to understand

4269: because its slow rate of dynamical evolution (if undisturbed by tides)

4270: means that it may retain a memory of its initial state.

4271: 

4272: 

4273: \subsubsection{\label{sub:Velocity-moments}Velocity moments.}

4274: 

4275: Integrating the collisionless Boltzmann equation (CBE) (2) over velocity

4276: yields the Jeans equations\begin{equation}

4277: \frac{\partial\overline{v}_{j}}{\partial t}+\overline{v}_{i}\frac{\partial\overline{v}_{j}}{\partial x_{i}}=-\frac{\partial\Phi}{\partial x_{j}}-\frac{1}{\nu}\frac{\partial(\nu\sigma_{ij}^{2})}{\partial x_{i}}\end{equation}

4278: where \begin{equation}

4279: \nu\equiv\int fd^{3}v\end{equation}

4280: is the mass density

4281:  and \begin{equation}

4282: v_{i}\equiv\frac{1}{v}\int f\, v_{i}d^{3}v\end{equation}

4283:  are the three velocity moments, and\begin{equation}

4284: -v\sigma_{ij}^{2}\equiv-v(\overline{v_{i}}\,\overline{v_{j}}-\overline{v_{i}v_{j}})\end{equation}

4285:  is the velocity dispersion stress tensor with \begin{equation}

4286: \overline{v_{i}v_{j}}\equiv\frac{1}{v}\int v_{i}v_{j}fd^{3}v\end{equation}

4287: As mentioned, there is no equation of state

4288: to relate motions (through tensor $\boldsymbol{\sigma}^{2}$)

4289: to mass density $\nu$. Integrating higher moments $v_{i}v_{k}$ of

4290: the CBE over velocity does not make this link, it simply introduces

4291: an unspecified third-order tensor $\overline{v_{i}v_{j}v_{k}}$. The

4292: only route forward is to truncate the moment sequence, for example

4293: by making hopefully reasonable, but nonetheless non-unique, assumptions

4294: on $\boldsymbol{\sigma}^{2}$. This approach was first applied to

4295: galaxy formation by \citet{Larson69}, who truncated and then integrated

4296: numerically the stellar and gas dynamical moment equations to model

4297: observed properties of Egs.

4298: 

4299: For example, a non-rotating galaxy with rotationally invariant densities

4300: and velocities has $\overline{v_{\theta}^{2}(r)}=\overline{v_{\phi}^{2}}(r)$.

4301: The anisotropy of tensor $\boldsymbol{\sigma}^{2}$ is described by 

4302: 

4303: \begin{equation}

4304: \beta(r)=1-\overline{v_{\theta}^{2}(r)}/\overline{v_{\rm{r}}^{2}}(r)\end{equation}

4305: with $\beta=-\infty,\,0,$ and 1 for orbits that are circular, isotropic,

4306: and radial, respectively. Even with this simplifying assumption,

4307: we unfortunately have only five equations to solve for six variables:

4308: $\nu$, $\overline{v_{\rm{r}}}$, $\overline{v_{\theta}}$, $\Phi$,

4309: $\overline{v_{\rm{r}}^{2}}$, and $\beta$ (mass density, two independent

4310: mean velocities, gravitational potential, velocity dispersion along

4311: radii in the galaxy, and anisotropy of the velocity dispersion). Next,

4312: we assume a spherically

4313: symmetrical Eg in a static configuration, i.e., without streaming

4314: motions. The Jeans equations then simplify to\begin{equation}

4315: \frac{1}{\nu}\frac{d(\nu\overline{v_{\rm{r}}^{2}})}{dr}+2\frac{\beta\overline{v_{\rm{r}}^{2}}}{r}=-\frac{d\Phi}{dr}\end{equation}

4316: Setting \begin{equation}

4317: \frac{\partial\Phi}{\partial r}=\frac{GM(r)}{r^{2}}\end{equation}

4318: recasts the radial Jeans equation into\begin{equation}

4319: v_{\rm{c}}^{2}=\frac{GM(r)}{r}=-\overline{v_{\rm{r}}^{2}}(\frac{d\ln\nu}{d\ln r}+\frac{d\ln\overline{v_{\rm{r}}^{2}}}{d\ln r}+2\beta)\end{equation}

4320: to solve for $M(r)$ \citep[see section 4.2 of][for details]{Binney87}.

4321: If the radial variations of mass density, velocity

4322: dispersion, and velocity anisotropy $\beta$ can be assumed or determined

4323: empirically by deprojecting measurements of luminous tracers such

4324: as planetary nebulae or globular clusters, we can recover the mass distribution. For example,

4325: \citet{Sargent78} assumed $\beta=0$ (i.e., isotropic velocity dispersion)

4326: in the central regions of Eg M87, and found from the inward rise in

4327: both luminosity and velocity function dispersion that $\Upsilon$ must increase toward

4328: the centre, indicating a supermassive black-hole. Subsequently, \citet{Binney82} relaxed

4329: the assumption $\beta=0$, and found that rising $\Upsilon$ is unnecessary

4330: if large ad hoc variations in $\beta(r)$ are allowed.

4331: 

4332: \subsubsection{Choice of velocity distribution function.}

4333: 

4334: Using only velocity moments does not ensure a viable solution of the

4335: Jeans equations (positive distribution function everywhere). Nor is any solution based

4336: on an assumed velocity dispersion tensor necessarily stable. An alternative

4337: approach to solving the CBE is to use prior knowledge of the distribution function. For

4338: a spherical galaxy, assume that the distribution function is both Maxwellian and `isothermal',

4339: namely an isotropic velocity dispersion that is independent of radius.

4340: Unfortunately, such a galaxy has infinite mass and radius.

4341: When a dynamical system can be truncated tidally, a `lowered

4342: Maxwellian' can be introduced. The distribution function of a lowered Maxwellian that truncates

4343: at positive energy $\varepsilon$ is\begin{equation}

4344: f_{\rm{K}}(\varepsilon)=\left\{ \begin{array}{cc}

4345: \rho(2\pi\sigma^{2})^{-3/2}(\exp[\varepsilon/\sigma^{2}]-1) & \varepsilon>0\\

4346: 0 & \varepsilon\le0\end{array}\right.\end{equation}

4347: \citet{Michie63} used (A.12) to model self-consistently the mass distributions

4348: of globular clusters, and \citet{King66} derived a model family that fit their

4349: SB profiles, although velocity function and proper motion measurements

4350: later showed anisotropic motions outside the cores.

4351: As discussed in section \ref{sub:Intrinsic-shape}, observations show

4352: that Egs are actually triaxial, so have anisotropic velocity dispersions.

4353: Egs also have structured centres, not the constant density core of

4354: King models. Hence, King models are overly restrictive

4355: and are not fully realized in Egs.

4356: 

4357: 

4358: \subsubsection{Integrals of motion.}

4359: 

4360: The CBE can also be solved in terms of integrals of motion, namely

4361: independent functions of phase-space $(\mathbf{x},\mathbf{v})$ that

4362: have constant value along any orbit. In a static potential (no dissipation),

4363: energy $E(\mathbf{x},\mathbf{v})=\frac{1}{2}v^{2}+\Phi(\mathbf{x})$

4364: is one and, if potential $\Phi$ is axisymmetric, the component of

4365: angular momentum around the rotation axis is too. In a spherical potential,

4366: angular momentum $\mathbf{L}$ yields three integrals of motion. If

4367: the potential is central, there is a fifth integral whose five-dimensional

4368: surface in phase space $\psi(\mathbf{x},\mathbf{v})=$constant intersects

4369: the 2D surface of constant $E$ and $\mathbf{L}$ to form a rosette-pattern

4370: orbit that is restricted to a five-dimensional region of phase space.

4371: In a $1/r$ potential, the rosette closes and the fifth integral is

4372: said to be isolating. Otherwise, the entire 2D surface of constant

4373: $E$ and $\mathbf{L}$ is covered eventually by the non-isolating

4374: integral.

4375: 

4376: The dimensionality (hence complexity) of solutions to the CBE reduces

4377: greatly with solutions that depend only on isolating

4378: integrals. Orbits are regular if they have at least as many isolating

4379: integrals as spatial dimensions. Realistic potentials have fewer isolating

4380: integrals, hence have irregular orbits. The Strong Jeans Theorem states

4381: that a distribution function composed of regular orbits is a function of just three isolating

4382: integrals, $f(E,I_{2},I_{3})$. 

4383: 

4384: For example, assuming that $E$ is the only isolating integral in

4385: a spherical galaxy is equivalent to assuming an isotropic velocity

4386: dispersion tensor, which we noted is overly restrictive. To accord

4387: fully with the Strong Jeans Theorem requires a distribution function that depends on

4388: two further integrals without explicit forms. So, one next considers

4389: distribution functions that depend on both $E$ and $\left|\mathbf{L}\right|$. Generally,

4390: the third integral is far harder to identify.

4391: 

4392: 

4393: \subsubsection{\label{sub:Stellar-orbits-and}Stellar orbits and the gravitational

4394: potential.}

4395: 

4396: A galaxy can also be described by

4397: the orbit families of an assumed potential that reproduce its inferred

4398: distribution function and stellar motions.

4399: 

4400: An axisymmetric potential usually has a third integral with

4401: $E$ and $L_{z}$. Some stars oscillate harmonically parallel

4402: to the cylindrical coordinate axes to form box orbits (Lissajous figures,

4403: which have no fixed sense of rotation and carry a star arbitrarily

4404: close to the centre of mass). At larger radii one finds loop orbits

4405: that are nearly circular and whose initial tangential velocity determines

4406: the elliptical annulus that confines the orbit. This resembles

4407: the annulus filled by an orbit in an axisymmetric potential as

4408: $L_{\rm{z}}$ is varied. Box orbits fill the phase space

4409: of more flattened axisymmetric systems. Orbital resonances,

4410: central density cusps, or central point masses change box orbits to

4411: boxlet orbits that tend to avoid small radii, or stochastic orbits

4412: that cover phase space chaotically when the central mass grows to $\sim1\%$

4413: of the system mass.

4414: 

4415: Realistic potentials and mass distributions thus have complex

4416: orbits to integrate numerically, but restricted potentials can

4417: be studied analytically. In particular, Stäckel (Eddington) potentials

4418: are the most general that separate in curvilinear coordinates, hence

4419: all orbits have three isolating integrals.

4420: Realistic galactic potentials (including those for triaxial figures)

4421: include both regular and irregular orbits, so Stäckel potentials are

4422: too restrictive. But their solutions are far more tractable because

4423: the moment equations now form a closed set. \citet{vandeVen03} discuss

4424: general solutions obtained with Stäckel potentials.

4425: 

4426: 

4427: \subsubsection{Spatial moments: tensor virial theorem.}

4428: 

4429: In \ref{sub:Velocity-moments} we obtained three Jeans velocity

4430: moments by multiplying the CBE by each velocity component

4431: and then integrating over all velocities. Now multiply

4432: each velocity moment by a \textit{position}

4433: coordinate and then integrate over all positions to form nine

4434: tensor virial theorem (tensor Virial Theorem) equations\begin{equation}

4435: \frac{1}{2}\frac{d^{2}I_{jk}}{dt^{2}}=2T_{jk}+\Pi_{jk}+W_{jk}\end{equation}

4436: with the moment of inertia tensor \begin{equation}

4437: I_{jk}=\int\rho x_{j}x_{k}d^{3}x\end{equation}

4438: partitioned into the potential energy tensor \begin{equation}

4439: W_{jk}=-\int\rho(x)x_{j}\frac{\partial\Phi}{\partial x_{k}}d^{3}x\end{equation}

4440:  and the kinetic energy tensor \begin{equation}

4441: K_{jk}=\frac{1}{2}\int\rho\overline{v_{j}v_{k}}d^{3}x=T_{jk}+\frac{1}{2}\Pi_{jk}\end{equation}

4442: that sums ordered $T$ and random $\Pi$ motions, respectively. A `cold'

4443: dynamical system is one where $W_{jk}$ and $T_{jk}$ balance, a `hot'

4444: system balances $W_{jk}$ and $\Pi_{jk}$. 

4445: The tensor Virial Theorem links the global spatio-kinematical properties of a galaxy.

4446: While spatial data are straightforward to obtain in projection, kinematics

4447: were limited until recently to samples along a PA's through

4448: narrow slits. Complete spatio-kinematical coverage over the inner

4449: part of an Eg is now attained in a single pointing of an integral-field

4450: spectrograph (IFS).

4451: 

4452: The most important insight from the tensor Virial Theorem follows when gravity collects

4453: matter from rest. As matter equilibrates, the

4454: tensor Virial Theorem tells us that half of the gravitational energy released causes

4455: motion within the potential well and the other half is dissipated

4456: to achieve binding energy $E_{b}=-E$ equal to the kinetic energy;

4457: the system has negative specific heat. For the MWg of mass $M_{\rm{g}}$

4458: and average rotational velocity $\Theta_{0}\approx200$ \kms.

4459: $E_{\rm{b}}=K\approx\frac{1}{2}M_{\rm{g}}\Theta_{0}^{2}$, hence

4460: $\approx2.5\times10^{-7}$ of its rest mass energy must have dissipated

4461: during its formation.

4462: 

4463: The tensor Virial Theorem can estimate the global $\Upsilon$ of

4464: a nonrotating spherical galaxy by averaging stellar

4465: velocity dispersion $\sigma(R)$ and SB $\sigma(R)$ maps 

4466: over the projected radius $R$ from

4467: IFS spectra\begin{equation}

4468: \widehat{I}=3\pi\int_{0}^{\infty}\sigma(r)\sigma^{2}(r)rdr,\end{equation}

4469: and by recognising that density follows from an Abel inversion integral

4470: of the measured radial variation of the SB. 

4471: With $\overline{I}$ an integral of $\sigma(R)$, the tensor Virial Theorem then gives

4472: \begin{equation}

4473: M/L\equiv\Upsilon=-\frac{2\widehat{I}}{\overline{I}}\end{equation}

4474: In section \ref{sec:Spheroidal-Galaxies}

4475: the tensor Virial Theorem constrains Eg 3D shapes by comparing radial

4476: trends of mean rotational/random motion with SB ellipticity.

4477: tensor Virial Theorem volume averaged properties require non-unique assumptions.

4478: 

4479: 

4480: \subsubsection{Violent relaxation.}

4481: 

4482: Statistical mechanical quantities that depend on randomization such

4483: as system temperature and entropy are seemingly undefinable in a collisionless

4484: stellar system. Yet, virialization implies past dissipation. Violent

4485: relaxation \citet{Lynden67} is the dissipator that erases most initial conditions by

4486: widening the global range of stellar energies (but perhaps not the

4487: energies of certain orbit families), independent of stellar mass.

4488: It arises from rapid fluctuations in the gravitational field during

4489: hierarchical mass buildup that scatters the stellar energies. 

4490: Simulations \citep[for example][]{Dekel05}

4491: show that the baryons approach virial equilibrium after a few dynamical

4492: times, namely a Maxwell distribution that is nearly isotropic near

4493: the centre\begin{equation}

4494: f(\varepsilon)=\frac{\rho}{(2\pi\sigma^{2})^{3/2}}\exp(\frac{\Psi-v^{2}/2}{\sigma^{2}}),\end{equation}

4495: with isothermal equation of state $p(\rho)=K\rho$ and with $\Psi$ the relative

4496: potential. Meanwhile, stars outside remain anisotropic,

4497: with many on radial orbits that originated in the mass buildup

4498: (see section \ref{sub:Dark-matter-content}).

4499: 

4500: In the CDM paradigm, DM relaxes to form the potential well. The baryons

4501: accrete into the well --- their dissipation characterized by parameter

4502: \begin{equation}

4503: \lambda=\widehat{J}\left|E\right|^{1/2}\frac{1}{GM^{3/2}}\end{equation}

4504: with $\widehat{J}\equiv J/M$, which

4505: sets \citep[for example][]{Dalcanton97} the SB of

4506: the baryonic disc. Figure \ref{fig:Abadi-et-al} shows the distribution

4507: of angular momentum from a $\Lambda$-CDM simulation \citep{Abadi03}, but

4508: present data are too sparse to constrain such models.

4509: 

4510: As discussed in section \ref{sec:Bulges-of-Spiral}, $\widehat{J}$

4511: of a disc spreads locally by collective

4512: motions from dynamical resonances in quasi-periodic orbits. Some DM

4513: distributions have minimal resonances hence maximal $\widehat{J}$

4514: at the disc rim. We cannot yet measure stellar kinematics at

4515: this SB, but can see if the stellar disc apears to truncate (section 5.1.2).

4516: 

4517: \begin{figure}

4518: \begin{centering}\includegraphics[scale=1.8]{figure18.pdf}\par\end{centering}

4519: \caption{\label{fig:Abadi-et-al}\citet{Abadi03}'s simulated formation of

4520: an Sa galaxy. The z-component of specific angular momentum within $R_{\rm{e}}$

4521: of the baryonic galaxy is plotted versus the specific binding

4522: energy. Inserts plot the distribution of angular momentum

4523: (normalized to a fraction of circular orbit value) of gas and stars,

4524: together with a possible decomposition into different components

4525: (labelled \dotted\ for thin disc, \broken\

4526: for thick disc, \full\ for spheroid). Despite photometric

4527: resemblance to Sa galaxy components, the simulated rotation curve

4528: is too centrally peaked. }

4529: \end{figure}

4530: 

4531: \subsubsection{\label{sub:Building-a-mass}Building a mass distribution and potential

4532: self-consistently from stellar orbits.}

4533: 

4534: To model self-consistently a galaxy, one combines orbit

4535: families of assumed $\Upsilon$ with a set of global weights

4536: to reproduce the distribution function as derived from the SB profile and the velocity function including

4537: asymmetries in absorption line profiles. Elucidating radial variations

4538: of $\Upsilon$ would be an important insight from a successful model,

4539: but current data are inadequate. At best, one posits a mass

4540: distribution with a spherical DM halo and stars whose average $\Upsilon$

4541: is consistent with the observed radial profiles of absorption line

4542: strengths and photometric colours. The distribution function is assumed to be steady

4543: state, but recall that galaxy shape may arise from both initial

4544: conditions and external perturbations. 

4545: To model an Eg one assumes inclination $i$ and then 

4546: \begin{itemize}

4547: \item 

4548: use either a constraining technique such as

4549: maximum entropy, or, fit to a functional form that has an analytical inversion, to

4550: deproject the mapped SB $\sigma(R,z)$ with the Abel integral

4551: \begin{equation}

4552: \rho(r,z)=-\frac{\Upsilon}{\pi}\int_{r}^{\infty}\frac{d\sigma(R,z)}{dR}\frac{dR}{\sqrt{R^{2}-r^{2}}}\end{equation}

4553: \item calculate the corresponding potential $\Phi(R,z)$ assuming constant

4554: $\Upsilon$;

4555: \item solve Jeans equations for mean velocities, then divides

4556: random and streaming motions;

4557: \item integrate space motions to get the velocity function and dispersion;

4558: \item compare to data and iterate to refine $\Upsilon$, the velocity

4559: division, and $i$. A valid model has positive distribution function throughout

4560: the galaxy phase space, while satisfying (\ref{eqn:poisson}).

4561: \end{itemize}

4562: Velocity dispersions along the galaxy minor axis are always observed

4563: to be smaller than predicted by this procedure, implying a third integral

4564: and/or triaxial shape (section \ref{sub:Intrinsic-shape}). Surface photometry

4565: finds non-elliptical isophotes for spheroidal galaxies. Some patterns

4566: are consistent with embedded discs, which may include dust and gas

4567: moving on inclined rings and counter-rotating compared to the stars.

4568: The cold ISM settles into stable orbits by differential (with radius)

4569: precession that collides clouds and dissipates their kinetic energy.

4570: The ISM may form a warped disc in an oblate potential or polar ring

4571: if prolate. An example is the dusty, but optically thin in near-IR, warped

4572: disc in Centaurus A (NGC 5128), studied recently with \textit{SST} \citep{Quillen06};

4573: based on transient tidal features (the warp and outer stellar shells),

4574: it may have arisen as a small gas-rich satellite merged $\sim0.1$

4575: Gyr ago.

4576: 

4577: Studies that assume a fixed, or slowly varying, potential from

4578: a mass model $\rho(r)$ use the orbit summing method of \citet{Schwarzschild79}.

4579: Many orbits are computed for many oscillations in the specified

4580: potential; most are tubes or boxes without net rotation. By noting

4581: regions within an equipotential surface that some orbits never traverse,

4582: Schwarzschild found three integrals of motion for most orbits in his triaxial model potential.

4583: Next he summed orbits with positive

4584: weights to build the assumed $\rho(r)$. He found the weights by linear

4585: programming, but now maximum entropy and the \citet{Lucy74}

4586: or related stack-smooth-iterate \citep{Bak00} algorithms are used.

4587: Methods such as `dithered' initial conditions for orbit integration

4588: \citep{Cappellari06} ensure the smooth distribution functions required to fit modern

4589: data; today over 400,000 orbits are summed with weights selected to

4590: reproduce the stellar density and velocity function.

4591: 

4592: Different distribution functions are consistent with a given mass model because

4593: the velocity function is degenerate with galaxy inclination unless additional features

4594: such as a stellar disc, ionized gas \citep[for example][]{Krajonovic05},

4595: neutral gas \citep{Morganti06}, or dust lanes \citep{Quillen06}

4596: also constrain inclination. If the mass distribution and potential

4597: are constrained by x-ray data, the orbital anisotropy is no longer

4598: degenerate. Today Jeans models are often used because of their transparency,

4599: uniqueness, and efficiency to generate an ensemble of noise-free distribution functions

4600: across a grid of structural parameters for Monte Carlo error analysis

4601: of the more sophisticated but non-unique Schwarzschild technique \citep[for example][]{Krajonovic05}.

4602: 

4603: 

4604: \section{A Brief Review of Hierarchical Galaxy Formation}

4605: 

4606: Hierarchical buildup from small to large mass develops cosmic structure

4607: from the early linear phase explored in the spectrum of temperature

4608: fluctuations in the CMBR into the non-linear regime. In particular

4609: \citep[for a comprehensive review, see][]{Baugh06}, $\Lambda$-CDM matches

4610: the power spectrum of the fluctuations

4611: as mapped by WMAP and terrestrial

4612: instruments, the Hubble diagram for high redshift

4613: SNe Ia, and the galaxy correlation function in recent wide-angle galaxy

4614: surveys. \textit{Direct} study of underlying DM halos

4615: is impossible. Instead, we study the baryonic component that dissipates

4616: energy within the DM halo, forms stars, and leaves hot and cool gas.

4617: Key for a researcher interested only in the halo is to determine to

4618: what extent light traces DM. Appendix A introduces some of the galactic astrophysics that

4619: intrudes.

4620: 

4621: Galaxy and cluster build up in the $\Lambda$-CDM

4622: framework has been explored by simulations that

4623: follow up to $10^{9}$ DM `particles' from post-recombination

4624: density perturbations to the present epoch \citep[for example][]{Baugh06}.

4625: Volumes $\sim500$ Mpc on a side have been

4626: simulated with DM particles as small as $10^{8}$ \Ms.

4627: Because DM is cold and non-interacting,

4628: gravitational N-body simulations can track its concentration into halos.

4629: In this way

4630: galaxy masses as small as $10^{9}$ \Ms\  are followed,

4631: unfortunately not the dwarf end of the galaxy mass function.

4632: Alternatively, galaxy build up can be followed in a statistical sense

4633: through semi-analytic prescriptions founded on the work of \citet{Press74}.

4634: The consensus in comparing both methods is that the predictions

4635: of $\Lambda$-CDM simulations are robust.

4636: However, study of baryonic matter within CDM halos requires

4637: accurate models of galaxy astrophysics over the mass range

4638: encompassed locally by a box $\sim12$ Mpc on a side. For example,

4639: almost all SF seen at high redshifts comes from a cooling flow into

4640: CDM halos, yet its non-linear effects on

4641: the surrounding ISM (`feedback', see section \ref{sub:Feedback})

4642: is well beyond present simulations.

4643: 

4644: A successful model must also reproduce the fate of baryons as galaxies merge.

4645: \citet{Helly03} compares the two tactics in use:

4646: 1) solve simultaneously the equations of hydrodynamics for

4647: gaseous baryons during the N-body simulation of the DM halo \citep[for example][]{Katz92,Steinmetz99,Cen00,Pearce01}.

4648: Naturally, at some point considerable sub-grid astrophysics must be

4649: parametrized. 2) graft this parametrized astrophysics onto statistical

4650: semi-analytic models \citep[for example][]{Cole00}.

4651: We are still far from fundamental treatment of the gaseous baryons.

4652: 

4653: The starting point as the DM halo forms is to cool and settle baryonic

4654: gas into a SF disc. An Eg forms when two near-equal mass disc galaxies

4655: merge. It can later grow a new disc to form a large-bulge Sg only if hot

4656: halo gas remaining from the merger can cool and settle. The scenario

4657: reproduces the scale of early-type galaxies by forming galaxies in

4658: DM halos where baryon cooling is efficient: objects lighter than

4659: $10^{12}$ \Ms\  form galaxies, those heavier form clusters. A persistent

4660: problem is that this scenario concentrates baryons too centrally,

4661: the so-called `angular momentum catastrophe'. DM halos acquire

4662: much of their angular momentum in mergers from the orbits of the progenitor

4663: galaxies. The baryonic gas, in cooling early toward the centres of

4664: the DM halos, does not gain this angular momentum \citep[for example][]{Steinmetz99,Navarro94}.

4665: $\Lambda$-CDM simulations therefore predict very structured halos

4666: and a central cusp of DM.

4667: 

4668: Despite these difficulties, simulations have provided robust results.

4669: Specifically, larger DM halos tend to assemble at lower redshifts

4670: (i.e., more recently) than smaller ones, and in general DM halos

4671: form more rapidly in denser than sparser environments \citep{Baugh06}.

4672: 

4673: The situation for baryons, particularly baryons now in

4674: stars, is murkier. Early simulations predicted that they would

4675: indeed follow DM through a process of adiabatic compression. More massive

4676: Egs are predicted to have younger luminosity-weighted mean ages than

4677: lower mass Egs, and those in clusters were predicted to be older in

4678: the mean than those in sparser regions. Egs of given mass

4679: are predicted to scatter significantly in luminosity-weighted mean

4680: age because of the stochastic nature of mergers \citep[for example][]{Kauffman98}.

4681: However, as we discuss in section \ref{sec:Spheroidal-Galaxies} and section \ref{sec:Bulges-of-Spiral},

4682: the inferred mean ages of Egs indicate that low mass Egs typically

4683: contain younger stars than massive Egs. This anti-hierarchical

4684: behaviour (called `down-sizing') led to new prescriptions for

4685: cooling and feedback (see section \ref{sub:Feedback}) related to SF and

4686: active galactic nuclei (AGN). For example, \citet{deLucia05} suppress

4687: cooling of massive Egs that ends SF early on, while

4688: \citet{Croton06} and \citet{Bower06}

4689: use AGN for mass-dependent feedback to inhibit hierarchical mass

4690: build up of baryons.

4691: 

4692: In summary, observations are now connecting with models but both

4693: are still too uncertain to establish

4694: firmly the viability of the $\Lambda$-CDM scenario in the non-linear

4695: regime.

4696: 

4697: \section*{References}

4698: \begin{harvard}

4699: \bibitem[Abadi \etal(2003)]{Abadi03}Abadi  M  G, Navarro J F, Steinmetz

4700: M and Eke V R 2003 \apj \textbf{597} 21--34

4701: 

4702: \bibitem[Abadi \etal(1999)]{Abadi99}Abadi M G, Moore B and Bower

4703: R G 1999 \mnras \textbf{308} 947--54

4704: 

4705: \bibitem[Alcock \etal(2000)]{Alcock00}Alcock C \etal\  2000 \apj

4706: \textbf{542} 281--307 

4707: 

4708: \bibitem[Antia and Basu(2006)]{Antia06}Antia H M and Basu S 2006

4709: \apj \textbf{644} 1292--8

4710: 

4711: \bibitem[Ashman and Zepf(1992)]{Ashman92}Ashman K E and Zepf S E

4712: 1992 \apj \textbf{384} 50--61

4713: 

4714: \bibitem[Asplund \etal(2005)]{Asplund05a}Asplund M, Grevesse N, Sauval

4715: A J, Allende Prieto C and Kiselman D 2005 \aap 435 339--40

4716: 

4717: \bibitem[Athanassoula(2003)]{Athanassoula03a}Athanassoula E 2003 \mnras \textbf{341} 1179--98

4718: 

4719: \bibitem[Athanassoula and Bosma(2003)]{Athanassoula03b}Athanassoula E and Bosma A 2003 

4720: \textit{Astophy. Space Sci.} \textbf{284} 491--4

4721: 

4722: \bibitem[Aubourg \etal(1993)]{Aubourg93}Aubourg E \etal\ 1993 \nat \textbf{365} 623--5

4723: 

4724: \bibitem[de Avillez(2000)]{deAvillez00}de Avillez M  A  2000 \mnras

4725: \textbf{315} 479--97

4726: 

4727: \bibitem[Azzaro \etal(2007)]{Azzaro06}Azzaro  M  \etal\  2007 \mnras \textbf{376} L43--7

4728: 

4729: \bibitem[Babusiaux and Gilmore(2005)]{Babusiaux05}Babusiaux  C  and

4730: Gilmore  G  2005 \mnras \textbf{358} 1309--19

4731: 

4732: \bibitem[Bahcall \etal(2005)]{Bahcall05}Bahcall J N, Basu S and Serenelii

4733: A M 2005 \apj \textbf{631} 1281--5

4734: 

4735: \bibitem[Bak and Statler(2000)]{Bak00}Bak  J  and Statler  T  S 2000

4736: \aj \textbf{120} 110--22

4737: 

4738: \bibitem[Baldry \etal(2004)]{Baldry04}Baldry  I K, Balogh M L, Bower R, Glazebrook

4739: K and Nichol R C 2004

4740: \textit{AIP Conf. Proc.} vol 743 (Melville, NY: AIP) p 106--19 

4741: (\textit{Preprint} astro-ph/0410603)

4742: 

4743: \bibitem[Balogh \etal(1999)]{Balogh99}Balogh M L, Morris S L, Yee H K C, Carlberg R G and 

4744: Ellingson E 1999 \apj \textbf{527} 54--79

4745: 

4746: \bibitem[Barbuy(1999)]{Barbuy99}Barbuy B 1999 \textit{Astrophy. Space

4747: Sci.} \textbf{265} 319--26

4748: 

4749: \bibitem[Bartelmann and Schneider(2001)]{Bartelmann01}Bartelmann

4750:  M  and Schneider  P  2001 \textit{Phys Reports} \textbf{340} 291

4751: 

4752: \bibitem[Barth(2007)]{Barth07}Barth A J 2007 \aj \textbf{133} 1085--91 

4753: 

4754: \bibitem[Baugh(2006)]{Baugh06}Baugh C M 2006 \RPP \textbf{69} 3101--56

4755: 

4756: \bibitem[Bautz and Arabadjis(2003)]{Bautz03}Bautz  M  W and Arabadjis

4757:  J  S  2003 

4758: \textit{Clusters of Galaxies: Probes of 

4759: Cosmological Structure and Galaxy Evolution} 

4760: (\textit{Carnegie Observatories Astrophysics Series} vol 3)

4761: ed J S  Mulchaey, A  Dressler and A Oemler (Cambridge: Cambridge Univ. Press)

4762: (\textit{Preprint} astro-ph/0303313)

4763: 

4764: \bibitem[Beaton \etal(2007)]{Beaton06}Beaton  R L \etal\  2007 \apj \textbf{658} L91--4

4765: 

4766: \bibitem[Bekki and Chiba(2000)]{Bekki00}Bekki K  and Chiba  M  2000

4767: \apj \textbf{534} L89--92

4768: 

4769: \bibitem[Bekki \etal(2005)]{Bekki05}Bekki K, Koribalski B S and Kilborn V A 2005 \mnras

4770: \textbf{363} L21--5

4771: 

4772: \bibitem[Bell and de Jong(2001)]{Bell01}Bell  E  F  and de Jong  R

4773:  S  2001 \apj \textbf{550} 212--29

4774: 

4775: \bibitem[Belokurov \etal(2006)]{Belokurov06}Belokurov V \etal\ 

4776: 2006 \apj \textbf{642} L137--40

4777: 

4778: \bibitem[Bender \etal(2005)]{Bender05}Bender  R  \etal\  2005 \apj

4779: \textbf{631} 280--300

4780: 

4781: \bibitem[Bennett \etal(2005)]{Bennett05}Bennett D P, Becker A C and

4782: Tomaney A 2005 \apj \textbf{631} 301--11

4783: 

4784: \bibitem[Benson \etal(2003)]{Benson03}Benson A J, Bower R G, Frenk

4785: C S, Lacey C G, Baugh C M and Cole S 2003 \apj \textbf{599} 38--49 

4786: 

4787: \bibitem[Berentzen \etal(2007)]{Berentzen07}

4788: Berentzen I, Shlosman I, Martinez-Valpuesta I and Heller C 2007 (\textit{Preprint} astro-ph/0703028)

4789: 

4790: \bibitem[Bertin and Lin(1996)]{Bertin96}Bertin G and Lin C C 1996 

4791: \textit{Spiral Structure in Galaxies} (Cambridge, MA: MIT Press)

4792: 

4793: \bibitem[Bertola and Capaccioli(1975)]{Bertola75}Bertola F and Capaccioli

4794: M 1975 \apj \textbf{200} 439--45

4795: 

4796: \bibitem[Bessell \etal(1998)]{Bessell98}Bessell M S, Castelli F and Plez B 1998 \aap \textbf{333} 231--50

4797: 

4798: \bibitem[Bicknell \etal(2000)]{Bicknell00}Bicknell  G, Sutherland

4799: R S, van Breugel, W J M, Dopita M A, Dey A and Miley G K  2000 \apj

4800: \textbf{540} 678--86

4801: 

4802: \bibitem[Binggelli \etal(1985)]{Bingelli85}Binggeli B, Sandage A

4803: and Tammann G A 1985 \aj \textbf{90} 1681--759 

4804: 

4805: \bibitem[Binney(1978)]{Binney78}Binney J 1978 \mnras \textbf{183}

4806: 501--14

4807: 

4808: \bibitem[Binney and Mamon(1982)]{Binney82}Binney J J and Mamon G

4809: A 1982 \mnras \textbf{200} 361--75

4810: 

4811: \bibitem[Binney and Tremaine(1987)]{Binney87}Binney J and Tremaine

4812: S 1987 \textit{Gravitational Dynamics} (Princeton: Princeton University

4813: Press)

4814: 

4815: \bibitem[Binney and Merrifield(1998)]{Binney98}Binney J and Merrifield

4816: M 1998 \textit{Galactic Astronomy} (Princeton: Princeton University

4817: Press)

4818: 

4819: \bibitem[Binney(2005)]{Binney05}Binney  J  2005 \mnras \textbf{363}

4820: 937--42

4821: 

4822: \bibitem[Bissantz and Gerhard(2002)]{Bissantz02}Bissantz N  and Gerhard

4823:  O  2002 \mnras \textbf{330} 591--608

4824: 

4825: \bibitem[Blackwell and Lynas-Gray(1998)]{Blackwell98}Blackwell  D

4826:  E and Lynas-Gray  A  E  1998 \aap \textbf{129} S505--15

4827: 

4828: \bibitem[Bland-Hawthorn \etal(1997)]{Bland-Hawthorn97}Bland-Hawthorn

4829:  J, Freeman K C and Quinn P J 1997 \apj \textbf{490} 143--55

4830: 

4831: \bibitem[Bland-Hawthorn and Freeman(2004)]{Bland-Hawthorn04}Bland-Hawthorn

4832:  J  and Freeman  K C  2004 \textit{The Fifth Workshop on Galactic Chemodynamics}

4833: \textit{Publ. Astro. Soc. Australia} \textbf{21} pp~110--20

4834: 

4835: \bibitem[Bland-Hawthorn \etal(2005)]{BH05}Bland-Hawthorn  J, Vlar\'e

4836: M, Freeman K C and Draine B T 2005 \apj \textbf{629} 239--49

4837: 

4838: \bibitem[Blitz \etal(2006)]{Blitz06}Blitz L, Fukui Y, Kawamura A,

4839: Leroy A, Mizuno N and Rosolowsky E 2006 \textit{Preprint} astro-ph/0602600

4840: 

4841: \bibitem[Borkova and Marsakov(2003)]{Bork03}Borkova T V and Marsakov V A 2003

4842: \aap \textbf{398} 133--9

4843: 

4844: \bibitem[Bosma(1981)]{Bosma81}Bosma A  1981 \aj \textbf{86} 1825--46

4845: 

4846: \bibitem[Bothun \etal(1997)]{Bothun97}Bothun G, Impey C and McGaugh

4847: S 1997 \pasp \textbf{109} 745--58

4848: 

4849: \bibitem[Bottema(1993)]{Bottema93}Bottema  R  1993 \aap \textbf{275}

4850: 16--36

4851: 

4852: \bibitem[Bournaud \etal(2005)]{Bournaud05}Bournaud F, Combes F and Semelin B 2005 

4853: \mnras \textbf{364} L18--22

4854: 

4855: \bibitem[Bower \etal(2006)]{Bower06}Bower R G, Benson A J, Malbon

4856: R, Helly J C, Frenk C S, Baugh C M, Cole S and Lacey C G 2006 \mnras 

4857: \textbf{370} 645--55 

4858: 

4859: \bibitem[Brinchmann \etal(2004)]{Brinchmann04}Brinchmann J, Charlot

4860: S, White S D M, Tremonti C, Kauffmann G, Heckman T and Brinkmann J

4861: 2004 \mnras \textbf{351} 1151--79

4862: 

4863: \bibitem[Broeils(1992)]{Broeils92}Broeils  A  H  1992 \aap \textbf{256}

4864: 19--32

4865: 

4866: \bibitem[Brown \etal(2006)]{Brown06}Brown T M, Smith E, Guhathakurta

4867: P, Rich R M, Ferguson H C, Renzini A, Sweigart A V and Kimble R A

4868: 2006 \apj \textbf{636} L89--92

4869: 

4870: \bibitem[Bruzual and Charlot(2003)]{Bruzual03}Bruzual G  and Charlot

4871:  S  2003 \mnras \textbf{344} 1000--34

4872: 

4873: \bibitem[Burstein(1979)]{Burstein79}Burstein D 1979 \apj \textbf{234}

4874: 829--36

4875: 

4876: \bibitem[Buta and Combes(1996)]{Buta96}Buta R and Combes F 1996

4877: \textit{Fundamentals of Cosmic Physics} \textbf{17} 95--281

4878: 

4879: \bibitem[Butcher and Oemler(1978)]{Butcher78}Butcher  H  and Oemler

4880:  A  1978 \apj \textbf{226} 559--65

4881: 

4882: \bibitem[Caldwell(1983)]{Caldwell83}Caldwell N 1983 \aj \textbf{88}

4883: 804--12

4884: 

4885: \bibitem[Caldwell \etal(1993)]{Caldwell93}Caldwell N, Rose J A, Sharples R M,

4886: Ellis R S and Bower R G 1993 \aj \textbf{106} 473--92

4887: 

4888: \bibitem[Caldwell and Rose(1997)]{Caldwell97}Caldwell N C and Rose J A 1997

4889: \aj 492--520

4890: 

4891: \bibitem[Caldwell \etal(1999)]{Caldwell99}Caldwell  N,  Rose  J  A

4892:  and Dendy  K 1999 \aj \textbf{117} 140--56

4893: 

4894: \bibitem[Caldwell \etal(2003)]{Caldwell03}Caldwell N,  Rose  J  A

4895:  and Concannon  K  2003 \aj \textbf{125} 2891--926

4896: 

4897: \bibitem[Capacciolo \etal(1990)]{Capaccioli90}Capaccioli M, Held

4898: E V, Lorenz H and Vietri M1990 \aj \textbf{99} 1813--22

4899: 

4900: \bibitem[Capaccioli \etal(1991)]{Capaccioli91}Capaccioli M, Vietri

4901: M, Held E V and Lorenz H 1991 \apj \textbf{371} 535--40

4902: 

4903: \bibitem[Cappellari \etal(2005)]{Cappellari05}Cappellari M \etal\ 

4904: 2005 \textit{Preprint} astro-ph/0509470

4905: 

4906: \bibitem[Cappellari \etal(2006)]{Cappellari06}Cappellari M  \etal\ 

4907: 2006 \mnras \textbf{366} 1126--50

4908: 

4909: \bibitem[Cardiel \etal(2003)]{Cardiel03}Cardiel

4910: N, Gorgas J, Sanchez-Blazquez P, Cenarro A J, Pedraz S, Bruzual G

4911: and Klement J 2003 \aap \textbf{409} 511--22 

4912: 

4913: \bibitem[Carignan \etal(2006)]{Carignan06}Carignan  C, Chemin L,

4914: Huchtmeier W K and Lockman F J 2006 \apj \textbf{641} L109--12

4915: 

4916: \bibitem[Carney \etal(1996)]{Carney96}Carney  B  W, Laird J B, Latham

4917: D W and Aguilar L A 1996 \aj \textbf{112} 668--92

4918: 

4919: \bibitem[Carrera \etal(2002)]{Carrerra02}Carrera R, Aparicio A, Martinez-Delgado

4920: D and Alonso-Garcia J 2002 \aj \textbf{123} 3199--209 

4921: 

4922: \bibitem[Carollo(2005)]{Corollo03}Carollo C M 2005

4923: \textit{Co-evolution of black holes and galaxies} 

4924: (\textit{Carnegie Observatory Astrophysics Series} vol 1)  ed L C Ho

4925: (Cambridge: Cambridge University Press) pp~231--47

4926: 

4927: \bibitem[Cattaneo \etal(2007)]{Cattaneo07}Cattaneo  A, Blaizot J,

4928: Weinberg D H, Colombi S, Dave R, Devriendt J, Guiderdoni B, Katz N

4929: and Keres D 2007 \mnras \textbf{377} 63--76

4930: 

4931: \bibitem[Cattaneo \etal(2006)]{Cattaneo06}Cattaneo A, Dekel A,

4932: Devriendt J, Guiderdoni B and Blaizot J 2006 \mnras \textbf{370} 1651--65

4933: 

4934: \bibitem[Cecil \etal(2001)]{Cecil01}Cecil G., Bland-Hawthorn J, Veilleux

4935: S and Filippenko A V 2001 \apj \textbf{555} 338--55

4936: 

4937: \bibitem[Cen and Ostriker(2000)]{Cen00}Cen R and Ostriker J P 2000

4938: \apj \textbf{538} 83--91

4939: 

4940: \bibitem[Cervi\~no \etal(2002)]{Cervino02}Cervi\~no M, Mas-Hesse

4941: J M and Kunth D 2002 \aap \textbf{392} 19--31

4942: 

4943: \bibitem[Chapman \etal(2006)]{Chapman06}Chapman S, Ibata R, Lewis

4944: G F, Ferguson A M N, Irwin N, McConnachie A and Tavnir N 2006 \apj \textbf{653} 255--66

4945: 

4946: \bibitem[Chiba and Beers(2001)]{Chiba01}Chiba T  and Beers  T 2001

4947: \apj \textbf{549} 325--36

4948: 

4949: \bibitem[Clowe \etal(2006)]{Clowe06}Clowe D, Bradac M, Gonzalez A

4950: H, Markevitch M, Randall S W, Jones C and Zaritsky D 2006 \apj \textbf{648}

4951: L109--11 

4952: 

4953: \bibitem[Coelho \etal(2005)]{Coelho05}Coelho P, Barbuy B, Melendez J, 

4954: Schiabon R P and Castilho B 2005 \aap \textbf{443} 735--46

4955: 

4956: \bibitem[Cole, Baugh and Frenk(2000)]{Cole00}Cole S, Lacey C G, Baugh

4957: C M and Frenk C S 2000 \mnras \textbf{319} 168--204

4958: 

4959: \bibitem[Colless \etal(2001)]{Colless02}Colless M \etal\  2001 \mnras

4960: \textbf{328} 1039--63

4961: 

4962: \bibitem[Cooper \etal(2006)]{Cooper06}Cooper M C \etal\  2006 \mnras

4963: \textbf{370} 198--212

4964: 

4965: \bibitem[Courteau \etal(1996)]{Courteau96}Courteau S, de jong R S

4966: and Broeils A H 1996 \apj \textbf{457} L73--6

4967: 

4968: \bibitem[Courteau and Rix(1999)]{Courteau99}Courteau S and Rix H

4969:  1999 \apj \textbf{513} 561--71

4970: 

4971: \bibitem[Courteau \etal(2003)]{Courteau03}Courteau S, Andersen D

4972: R, Bershady M A, MacArthur L A and Rix H-W 2003 \apj \textbf{594}

4973: 208--24

4974: 

4975: \bibitem[C\^ot\'e \etal(2000)]{Cote00}C\^ot\'e  S, Carignan C and

4976: Freeman K C 2000 \aj \textbf{120} 3027--59

4977: 

4978: \bibitem[Cox \etal(2004)]{Cox04}Cox T J, Primack J, Jonsson P and

4979: Somerville R S 2004 \apj \textbf{607} L87--90

4980: 

4981: \bibitem[Cox(2005)]{Cox05}Cox  D  P  2005 \araa \textbf{43} 337--86

4982: 

4983: \bibitem[Cr\'ez\'e \etal(1998)]{Creze98}Cr\'ez\'e M, Chereul E,

4984: Bienayme O and Pichon C 1998 \aap \textbf{329} 920--36

4985: 

4986: \bibitem[Croft \etal(2006)]{Croft06}Croft S \etal\  2006 \apj \textbf{647} 1040--55

4987: 

4988: \bibitem[Croton \etal(2006)]{Croton06}Croton D J, Springel V, White

4989: S D M, de Lucia G, Frenk C S, Gao L, Jenkins A, Kauffmann G, Navarro

4990: J F and Yoshida N 2006 \mnras \textbf{365} 11--28

4991: 

4992: \bibitem[Crowl \etal(2005)]{Crowl05}Crowl H H, Kenney J D P, van Gorkom J H and Vollmer B 2005

4993: \apj \textbf{130} 65--72

4994: 

4995: \bibitem[Dalcanton \etal(1997)]{Dalcanton97}Dalcanton J J , Spergel

4996: D N and Summers F J 1997 \apj \textbf{482} 659--76

4997: 

4998: \bibitem[Davies \etal(1983)]{Davies83}Davies R L, Efstathiou G, Fall

4999: S M, Illingworth G and Schechter P L 1983 \apj \textbf{266} 41--57

5000: 

5001: \bibitem[Davies and Birkinshaw(1988)]{Davies88}Davies R L and Birkinshaw

5002: M 1988 \apjs \textbf{68} 409--47

5003: 

5004: \bibitem[de Blok and McGaugh(1996)]{deBlok96}de Blok W J G and McGaugh

5005: S S 1996 \apj \textbf{469} L89--92

5006: 

5007: \bibitem[de Blok and McGaugh(1997)]{deBlok97}de Blok W  J  G  and

5008: McGaugh  S  S  1997 \mnras \textbf{290} 533--52

5009: 

5010: \bibitem[deGrijs \etal(2001)]{deGrijs01}de Grijs R, Kregel M and

5011: Wesson K H 2001 \mnras \textbf{324} 1074--86

5012: 

5013: \bibitem[de Lucia \etal(2005)]{deLucia05}de Lucia G, Springel V,

5014: White S D M, Croton D and Kauffmann G 2006 \mnras \textbf{366} 499--509 

5015: 

5016: \bibitem[de Silva \etal(2006)]{deSilva05}de Silva G M \etal\ 2006 \aj \textbf{131}

5017: 455-60

5018: 

5019: \bibitem[de Vaucouleurs(1948)]{deVau48}de Vaucouleurs G 1948 \textit{Ann.

5020: Astrophys.} \textbf{11} 247--87

5021: 

5022: \bibitem[Debattista \etal(2006)]{Debatt06}Debattista V P, Mayer L, Carollo C M, Moore B,

5023: Wadsley J and Quinn T 2006 \apj \textbf{645} 209--27

5024: 

5025: \bibitem[Dekel \etal(2005)]{Dekel05}Dekel A, Stoehr F, Mamon G A,

5026: Cox T J, Novak G S and Primack J R 2005 \textit{Nature} \textbf{437}

5027: 707--10

5028: 

5029: \bibitem[DiBenedetto(1998)]{DiBenedetto98}Di Benedetto G P 1998 \aap

5030: \textbf{339} 858--71

5031: 

5032: \bibitem[Dolphin(2002)]{Dolphin02}Dolphin A E 2002 \mnras \textbf{332}

5033: 91--108 

5034: 

5035: \bibitem[Dominguez \etal(1999)]{Dominguez99}Dominguez I, Chieffi

5036: A, Limongi M and Straniero O 1999 \apj \textbf{524} 226--41

5037: 

5038: \bibitem[Dopita \etal(2002)]{Dopita02}Dopita M A, Groves B, Sutherland

5039: R S, Binette L and Cecil G 2002 \apj \textbf{572} 753--61

5040: 

5041: \bibitem[Dopita(2005)]{Dopita05}Dopita M A 2005

5042: \textit{The Spectral Energy Distribution of Gas-Rich Galaxies: Confronting 

5043: Models with Data} ed Popescu C C and Tuffs R J (Melville, NY: AIP) 

5044: \textit{AIP Conf. Proc.} vol 761 p 203--22

5045: (\textit{Preprint} astro-ph/0502339)

5046: 

5047: \bibitem[Dopita \etal(2006)]{Dopita06}Dopita M A, Fischera A, Sutherland

5048: R S, Kewley L J, Tuffs R J, Popesch C C, van Breugel W, Groves B A

5049: and Leitherer C 2006 \apj \textbf{647} 244--55

5050: 

5051: \bibitem[Douglas \etal(2002)]{Douglas02}Douglas  N  G  \etal\ 2002

5052: \pasp \textbf{114} 1234--51

5053: 

5054: \bibitem[Dressler(1980)]{Dressler80}Dressler  A 1980 \apj \textbf{236}

5055: 351--65

5056: 

5057: \bibitem[Dressler(1986)]{Dressler86}Dressler A 1986 \apj \textbf{301} 35--43

5058: 

5059: \bibitem[Dressler(1997)]{Dressler97}Dressler A, Oemler A Jr, Couch W J,

5060: Smail I, Ellis R S, Barger A, Butcher H, Poggianti B M and Sharples R M

5061: 1997 \apj \textbf{490} 577--91

5062: 

5063: \bibitem[Dressler \etal(2004)]{Dressler04}Dressler  A, Oemler A Jr,

5064: Poggianti B M, Smail I, Trager S, Shectman S A, Couch W J and Ellis

5065: R S 2004 \apj \textbf{617} 867--78

5066: 

5067: \bibitem[Driver \etal(2007)]{Driver07}Driver S P, Popescu C, Tuffs R J, Liske J,

5068: Graham A, De Propris R and Allen P D 2007 \mnras\ in press

5069: 

5070: \bibitem[Dwek \etal(1995)]{Dwek95}Dwek E, Arendt R G, Hauser M G,

5071: Kelsall T, Lisse C M, Moseley S H, Silverberg R F, Sodroski T J and

5072: Weiland J L 1995 \apj \textbf{445} 716--30 

5073: 

5074: \bibitem[Eggen \etal(1962)]{Eggen62}Eggen  O J, Lynden-Bell D and

5075: Sandage  A  R  1962 \apj \textbf{136} 748--66

5076: 

5077: \bibitem[Eisenhauer \etal(2003)]{Eisenhauer03}Eisenhauer  F, Schoedel

5078: R, Genzel R, Ott T, Tecza M, Abuter R, Eckart A and Alexander A 2003

5079: \apj \textbf{597} L121--4

5080: 

5081: \bibitem[Elmegreen and Falgorone(1996)]{Elmegreen02}Elmegreen B  and

5082: Falgorone  E  1996  \apj \textbf{471} 816--21

5083: 

5084: \bibitem[Epchtein \etal(1994)]{Epchtein94}Epchtein N \etal\  1994

5085: \textit{Astr. Sp. Sci.} \textbf{217} 3--9

5086: 

5087: \bibitem[Erwin \etal(2005)]{Erwin05}Erwin P, Beckman J E and Pohlen M 2005

5088: \apj \textbf{626} L81--4

5089: 

5090: \bibitem[Eskridge \etal(2000)]{Eskridge00}Eskridge P B \etal 2000 \aj \textbf{119} 536--44

5091: 

5092: \bibitem[Evans(1999)]{Evans99}Evans N J 1999 \araa \textbf{37} 311--62

5093: 

5094: \bibitem[Evans and Belokurov(2002)]{Evans02}Evans N W and Belokurov

5095: V 2002 \apj \textbf{567} :119--23

5096: 

5097: \bibitem[Faber(1972)]{Faber72}Faber S M 1973 \apj \textbf{179} 731--54

5098: 

5099: \bibitem[Faber \etal(1985)]{Faber85}Faber S M, Friel E D, Burstein

5100: D and Gaskell C M 1985 \apjs \textbf{57} 711--41

5101: 

5102: \bibitem[Faber \etal(2005)]{Faber05}Faber S M \etal\ 2004 \textit(Preprint) astro-ph/0506044

5103: 

5104: \bibitem[Fabian(1994)]{Fabian94}Fabian  A  C  1994 \araa \textbf{32}

5105: 277--318

5106: 

5107: \bibitem[Fellhauer \etal(2006)]{Fellhauer06}Fellhauer M \etal\ 

5108: 2006 \apj \textbf{651} 167--73

5109: 

5110: \bibitem[Feltzing(2004)]{Feltzing04}Feltzing S, Bensby T and Lundstrom

5111: I 2003 \aap \textbf{397} L1--4

5112: 

5113: \bibitem[Ferguson(1989)]{Ferguson89}Ferguson H C 1989 \aj 98 367--418 

5114: 

5115: \bibitem[Ferrarese and Merritt(2000)]{Ferrarese00}Ferrarese  L  and

5116: Merritt D  2000 \apj \textbf{539} L9--12

5117: 

5118: \bibitem[Ferreras \etal(2006)]{Ferras06}Ferreras I, Pasquali A, de

5119: Carvalho R R, de la Rosa I G and Lahav O 2006 \mnras \textbf{370}

5120: 828--36

5121: 

5122: \bibitem[Ferreras \etal(2005)]{Ferreras05}Ferreras  I, Saha P and

5123: Williams  L I  R  2005 \apj \textbf{623} L5--8

5124: 

5125: \bibitem[Ferri{\`e}re(2001)]{Ferriere01}Ferri\`ere K M 2001 Rev.

5126: Mod. Phys. \textbf{73} 1031--66

5127: 

5128: \bibitem[Font \etal(2006)]{Font06}Font A, Johnston K, Guhathakurta R, Majewski S and Rich M

5129: 2006 \aj, \textbf{131} 1436

5130: 

5131: \bibitem[Fragile \etal(2004)]{Fragile04}Fragile P  C, Murray S D,

5132: Anninos P and van Breugel W 2004 \apj \textbf{604} 74--87

5133: 

5134: \bibitem[Fraternali and Binney(2006)]{Fraternali06}Fraternali F and Binney J J 2006 \mnras

5135: \textbf{366} 449--66

5136: 

5137: \bibitem[Freeman and Bland-Hawthorn(2002)]{Freeman02}Freeman K and

5138: Bland-Hawthorn  J  2002 \araa \textbf{40} 487--537

5139: 

5140: \bibitem[Freeman and McNamara(2006)]{Freeman06}Freeman K and McNamara G 2006

5141: \textit{In Search of Dark Matter} (Chichester: Praxis Publishing)

5142: 

5143: \bibitem[Friedli \etal(1994)]{Friedli94}Friedli D, Benz W and Kennicutt R 1994

5144: \apj \textbf{430} L105--8

5145: 

5146: \bibitem[Fuhrmann \etal(1997)]{Fuhrmann97}Fuhrmann  K, Pfeiffer M,

5147: Frank C, Reetz J and Gehren T 1997 \aap \textbf{323} 909--22

5148: 

5149: \bibitem[Gallart \etal(1999)]{Gallart99}Gallart  C, Freedman W L, Aparicio

5150: A, Bertelli G and Chiosi C 1999 \aj \textbf{118} 2245--61

5151: 

5152: \bibitem[Gallart \etal(2005)]{Gallart05}Gallart C, Zoccali M and

5153: Aparicio A 2005 \araa \textbf{43} 387--434

5154: 

5155: \bibitem[Garc\'ia-Ruiz \etal(2002)]{Garcia02}Garc\'ia-Ruiz  I, Sancisi

5156: R and Kuijken K 2002 \aap \textbf{394} 769--89

5157: 

5158: \bibitem[Gavazzi(1987)]{Gavazzi87}Gavazzi G  1987 \apj \textbf{320}

5159: 96--121

5160: 

5161: \bibitem[Gavazzi \etal(1995)]{Gavazzi95}Gavazzi G, Contursi A, Carrasco

5162: L, Boselli A, Kennicutt R, Scodeggio M and Jaffe W 1995 \aap \textbf{304}

5163: 325--340

5164: 

5165: \bibitem[Gebhardt \etal(2000)]{Gebhardt00}Gebhardt K  \etal\  2000

5166: \apj \textbf{539} L13--6

5167: 

5168: \bibitem[Geha \etal(2002)]{Geha02}Geha M, Guhathakurta P and van

5169: der Marel R P 2002 \aj \textbf{124} 3073--87

5170: 

5171: \bibitem[Ghez \etal(2004)]{Ghez04}Ghez  A  M, Salim S, Hornstein

5172: S D, Tanner A, Lu J R, Morris M, Becklin E E and Duchêne G 2005 \apj

5173: \textbf{620} 744--57

5174: 

5175: \bibitem[Gilmore and Reid(1983)]{Gilmore83}Gilmore G and Reid N 1983

5176: \mnras \textbf{202} 1025--47

5177: 

5178: \bibitem[Giovanelli and Haynes(1985)]{Giovanelli85}Giovanelli R and Haynes M P 1985

5179: \apj \textbf{292} 404--25

5180: 

5181: \bibitem[Glazebrook and Bland-Hawthorn(2001)]{Glazebrook01}Glazebrook

5182: K and Bland-Hawthron J 2001 \pasp \textbf{113} 197--214

5183: 

5184: \bibitem[Gonz\'alez Delgado \etal(1999)]{Gonzalez99}Gonz\'alez-Delgado

5185:  R  M  \etal\  1999 \apjs \textbf{125} 489--509

5186: 

5187: \bibitem[Gordon and Sorochenko(2003)]{Gordon03}Gordon M A and Sorochenko R L 2003 

5188: \textit{Radio Recombination Lines, Their Physics and Astronomical Applications} 

5189: (New York: Springer)

5190: 

5191: \bibitem[Gould \etal(1997)]{Gould97}Gould  A,  Bahcall J  N   and

5192: Flynn  C  1997 \apj \textbf{482} 913--8

5193: 

5194: \bibitem[Gould \etal(1998)]{Gould98}Gould  A,   Flynn  C  and Bahcall

5195:  J  N  1998 \apj \textbf{503} 798--808

5196: 

5197: \bibitem[Graham and Guzm\'an(2003)]{Graham03}Graham A  W  and Guzm\'an

5198:  R  2003 \aj \textbf{125} 2936--50

5199: 

5200: \bibitem[Graham(2005)]{Graham05}Graham A W \textit{Near-fields cosmology with dwarf elliptical 

5201: galaxies} (Cambridge: Cambridge U. Press) IAU Colloquium Proc. 198 pp~303--10

5202: 

5203: \bibitem[Gratton \etal(2003)]{Gratton03}Gratton R  G, Carretta E,

5204: Desidera S, Lucatello S, Mazzei P and Barbieri M 2003 \aap \textbf{406}

5205: 131--40

5206: 

5207: \bibitem[Grebel, Gallagher and Harbeck(2003)]{Grebel03}Grebel, Gallagher

5208: and Harbeck 2003 \aj \textbf{125} 1926--39

5209: 

5210: \bibitem[Grebel(2005)]{Grebel05}Grebel E K 2005 \textit{AIP Conf. Proc.}

5211: vol 752 (Melville, NY: AIP) p 161--74

5212: 

5213: \bibitem[Guhathakurta \etal(2006)]{Guhathakurta06}Guhathakurta  R,

5214: Rich M R, Reitzel D B, Cooper M C, Gilbert K M, Majewski S R, Ostheimer

5215: J C, Geha M C, Johnston K V and Patterson R J 2006 \aj \textbf{131}

5216: 2497--513

5217: 

5218: \bibitem[Gunn and Gott(1972)]{Gunn72}Gunn J  E and Gott  R 1972 \apj

5219: \textbf{176} 1--19

5220: 

5221: \bibitem[Gunn \etal(1981)]{Gunn81}Gunn J E, Stryker L L and Tinsley

5222: B M 1981 \apj \textbf{249} 48--67

5223: 

5224: \bibitem[Hafner \etal(2003)]{Hafner03}Hafner L M, Reynolds R J, Tufte

5225: S L, Madsen G J, Jaehnig K P and Percival J W 2003 \apjs \textbf{149}

5226: 405--22

5227: 

5228: \bibitem[H\"aring and Rix(2004)]{Haring04}H\"aring N and Rix H -W

5229:  2004 \apj \textbf{604} L89--92

5230: 

5231: \bibitem[Hartwick(2002)]{Hartwick02}Hartwick  F D A  2002 \apj \textbf{576}

5232: L29--32

5233: 

5234: \bibitem[Heckman \etal(2005)]{Heckman05}Heckman T M \etal\  2005

5235: \apj \textbf{619} L35--8

5236: 

5237: \bibitem[Helly \etal(2003)]{Helly03}Helly J C, Cole S, Frenk C S,

5238: Baugh C M, Benson A, Lacey C and Pearce F R 2003 \mnras \textbf{338}

5239: 913--925

5240: 

5241: \bibitem[Helmi \etal(2003)]{Helmi03}Helmi  A, White S D M and

5242: Springel V 2003 \mnras \textbf{339} 834--48

5243: 

5244: \bibitem[Henry \etal(1997)]{Henry97}Henry  T  J, Ianna P A, Kirkpatrick

5245: J D and Jahreiss H 1997 \aj \textbf{114} 388--95

5246: 

5247: \bibitem[Ho \etal(1997)]{Ho97}Ho L  C, Filippenko A V and Sargent

5248: W L W 1997 \apj \textbf{487} 568--78

5249: 

5250: \bibitem[Hoekstra  \etal(2004)]{Hoekstra04}Hoekstra H,   Yee  H  K

5251:  C   and Gladders  M D  2004 \apj \textbf{606} 67--77

5252: 

5253: \bibitem[Hunt and Malkan(1999)]{Hunt99}Hunt L K, and Malkan M A 1999 \apj

5254: \textbf{516} 660--671

5255: 

5256: \bibitem[Hunter and Toomre(1969)]{Hunter69}Hunter  C and Toomre A

5257:  1969 \apj \textbf{155} 747--76

5258: 

5259: \bibitem[Ibata \etal(2001a)]{Ibata01a}Ibata  K, Irwin M, Lewis G,

5260: Ferguson A and Tanvir N 2001 \apj \textbf{551} 294--311

5261: 

5262: \bibitem[Ibata \etal(2001b)]{Ibata01}Ibata  K, Irwin M, Lewis G,

5263: Ferguson A and Tanvir N 2001 \nat \textbf{412} 49--52

5264: 

5265: \bibitem[Ibata \etal(2003)]{Ibata03}Ibata  K, Irwin M J, Lewis G

5266: F, Ferguson A M N and Tanvir N 2003 \mnras \textbf{340} L21--7

5267: 

5268: \bibitem[Ibata \etal(2004)]{Ibata04}Ibata K, Chapman S, Ferguson A, Irwin M and Lewis G

5269: 2004 \mnras \textbf{351} 117--24

5270: 

5271: \bibitem[Ibata \etal(2007)]{Ibata07}Ibata K, Martin N F, Irwin M, Chapman S, Ferguson A M N,

5272: Lewis G F and McConnachie A W 2007 \textit{Preprint} astro-ph/0704.1318

5273: 

5274: \bibitem[Iben(1974)]{Iben74}Iben I 1974 \araa \textbf{12} 215--56

5275: 

5276: \bibitem[Iben(1991)]{Iben91}Iben I 1991 \apjs \textbf{76} 55--114

5277: 

5278: \bibitem[Iglesias and Rodgers(1996)]{Iglesias96}Iglesias C A and

5279: Rogers F J 1996 \apj \textbf{464} 943--53

5280: 

5281: \bibitem[Illingworth(1977)]{Illingworth77}Illingworth G 1977 \apj

5282: \textbf{218} L43--7

5283: 

5284: \bibitem[Impey and Bothun(1997)]{Impey97}Impey C and Bothun G 1997

5285: \araa \textbf{35} 267--307

5286: 

5287: \bibitem[Ivezic \etal(2004)]{Ivezic04}Ivezic  Z , Lupton R H, Schlegel

5288: D, Jonston D, Gunn J E, Knapp G R, Strauss M A and Rockosi C M 2004

5289: \textit{Satellites and Tidal Streams} (San Francisco: Astro. Soc. Pacific)

5290: \textit{ASP Conf.\ Proc.} vol 327 p~104

5291: 

5292: \bibitem[Jetzer and Novati(2004)]{Jetzer05}Jetzer Ph and Novati S

5293: C 2004 \textit{Preprint} astro-ph/0407209 

5294: 

5295: \bibitem[Jimenez \etal(2005)]{Jimenez05}Jimenez R, Panter B, Heavens

5296: A F and Verde L 2005 \mnras 356 495--501 

5297: 

5298: \bibitem[Jones \etal(2000)]{Jones00}Jones L,  Smail I and Couch W

5299:  2000 \apj \textbf{528} 118--22

5300: 

5301: \bibitem[J{\o}rgensen \etal(1996)]{Jorgensen96}J{\o}rgensen  I, Franx

5302: M and Kjaergaard P 1996 \mnras \textbf{280} 167--85

5303: 

5304: \bibitem[Kallivayalil, van der Marel and Alcock(2006b)]{Kallivayalil06b}

5305: Kallivayalil N, van der Marel R P and Alcock C 2006 \apj \textbf{652} 1213

5306: 

5307: \bibitem[Kallivayalil \etal(2006a)]{Kallivayalil06a}

5308: Kallivayalil N \etal 2006 \apj \textbf{648} 772

5309: 

5310: \bibitem[Kannappan \etal(2004)]{Kannappan04}

5311: Kannappan S J, Jansen R A and Barton E J 2004 \aj \textbf{127} 1371--85

5312: 

5313: \bibitem[Katz \etal(1992)]{Katz92}Katz N, Hernquist L and Weinberg

5314: D H 1992 \apj \textbf{399} L109--12

5315: 

5316: \bibitem[Kauffman and Charlot(1998)]{Kauffman98}Kauffman  G and Charlot

5317:  S  1998 \mnras \textbf{294} 705--17

5318: 

5319: \bibitem[Kauffman \etal(2003)]{Kauffman03}Kauffman G \etal\ 2003 \mnras \textbf{341}  33--53

5320: 

5321: \bibitem[Kenney and Koopmann(1999)]{Kenney99}Kenney  J  and Koopmann

5322: 1999 \aj \textbf{117} 181

5323: 

5324: \bibitem[Kennicutt \etal(1994)]{Kennicutt94}Kennicutt R C, Tamblyn,

5325: \& Congdon 1994 \apj \textbf{435} 22--36

5326: 

5327: \bibitem[Kennicutt(1998)]{Kennicutt98}Kennicutt R C 1998 \araa \textbf{36}

5328: 189--231

5329: 

5330: \bibitem[Kent(1986)]{Kent86}Kent  S M  1986 \aj \textbf{91} 1301--27

5331: 

5332: \bibitem[Keres \etal(2005)]{Keres05}Keres  D, Kaatz N, Weinberg D

5333: H and Dav\'e R 2005 \mnras \textbf{363} 2--28

5334: 

5335: \bibitem[King(1962)]{King62}King I R 1962 \aj \textbf{67} 471--85

5336: 

5337: \bibitem[King(1966)]{King66}King I R 1966 \aj \textbf{71} 64--75

5338: 

5339: \bibitem[Kinman \etal(2007)]{Kin07}Kinman T D, Cacciari C, Bragaglia A, Buzzoni A and Spagna A

5340: 2007 \mnras \textbf{375} 1381--98

5341:   

5342: \bibitem[Kleyna \etal(2001)]{Kleyna01}Kleyna  J  T, Wilkinson M I,

5343: Evans N W and Gilmore G 2001 \apj \textbf{563} L115--8

5344: 

5345: \bibitem[Kleyna \etal(2005)]{Kleyna05}Kleyna  J  T, Wilkinson M I,

5346: Evans N W and Gilmore G 2005 \apj \textbf{630} L141--4

5347: 

5348: \bibitem[Knapen(1999)]{Knapen99}Knapen, J H 1999 \textit{The Evolution of

5349: Galaxies on Cosmological Timescales} (San Francisco: Astro. Soc. Pacific)

5350: \textit{ASP Conf.\ Proc.} vol 187 72--87

5351: 

5352: \bibitem[Knapen \etal(2000)]{Knapen00}Knapen, J H, Shlosman, I, and Peletier,

5353: R F 2000 \apj \textbf{529} 93--100

5354: 

5355: \bibitem[Kobayashi and Arimoto(1999)]{Kobayashi99}Kobayashi C and

5356: Arimoto N 1999 \apj \textbf{527} 573--99

5357: 

5358: \bibitem[Koch \etal(2007)]{Koch06}Koch \etal 2007 \aj \textbf{133} 270--83

5359: 

5360: \bibitem[Koch and Grebel(2006)]{Koch05}Koch A and Grebel E K 2006

5361: \aj \textbf{131} 1405--15

5362: 

5363: \bibitem[Koopmann and Kenney(1998)]{Koopman98}Koopmann R A and Kenney J D P 1998 \apj 

5364: \textbf{497} L75--9

5365: 

5366: \bibitem[Kormendy(1982)]{Kormendy82}Kormendy J 1982 \textit{Morphology and Dynamics of 

5367: Galaxies, Proc. of the Twelfth Advanced Course} Saas-Fee Switzerland Observatoire de 

5368: Geneve pp~113--288

5369: 

5370: \bibitem[Kormendy(1985)]{Kormendy85}Kormendy J 1985 \apj \textbf{295}

5371: 73--9

5372: 

5373: \bibitem[Kormendy and Richstone(1995)]{Kormendy95}Kormendy  J and

5374: Richstone  D  1995 \araa \textbf{33} 581--624

5375: 

5376: \bibitem[Kormendy and Kennicutt(2004)]{Kormendy04}Kormendy  J  and

5377: Kennicutt  R  2004 \araa \textbf{42} 603--683

5378: 

5379: \bibitem[Kormendy and Fisher(2005)]{Kormendy05}

5380: Kormendy J and Fisher D B 2005, Rev. Mex. A. A. \textbf{23} 101--8

5381: 

5382: \bibitem[Kormendy \etal(2006)]{Kormendy06}

5383: Kormendy J, Cornell M E, Block D L, Knapen J H and Allard E L 2006 \apj \textbf{642} 765--74

5384: 

5385: \bibitem[Kozhurina-Platais \etal(1997)]{Kozhurina97}Kozhurina-Platais

5386: V, Demarque P, Platais I, Orosz J A and Barnes S 1997 \aj \textbf{113}

5387: 1045--56

5388: 

5389: \bibitem[Krajonovi\'c \etal(2005)]{Krajonovic05}Krajonovi\'c  D, Cappelari

5390: M, Emsellem E, McDermid R M and de Zeeuw P T 2005 \mnras \textbf{357}

5391: 1113--33

5392: 

5393: \bibitem[Krajonovi\'c \etal(2007)]{Krajonovic06}Krajonovi\'c D, Sharp

5394: R and Thatte N 2007 \mnras \textbf{374} 385--98

5395: 

5396: \bibitem[Kranz et al(2003)]{Kranz03}Kranz  T, Slyz A  and Rix  H-W

5397: 2003 \apj \textbf{586} 143--51

5398: 

5399: \bibitem[Kroupa(2001)]{Kroupa01}Kroupa  P 2001 \mnras \textbf{322}

5400: 231--46

5401: 

5402: \bibitem[Kroupa(2002)]{Kroupa02}Kroupa  P 2002 \mnras \textbf{330}

5403: 707--18

5404: 

5405: \bibitem[Kuijken and Rich(2002)]{Kuijken02}Kuijken K and Rich R M

5406: 2002 \aj \textbf{124} 2054--66 

5407: 

5408: \bibitem[Kuntschner(2000)]{Kuntschner00}Kuntschner  H 2000 \mnras

5409: \textbf{315} 184--208

5410: 

5411: \bibitem[Lambas \etal(1992)]{Lambas92}Lambas D G, Maddox S J and

5412: Loveday J 1992 \mnras \textbf{258} 404--14

5413: 

5414: \bibitem[Larson(1969)]{Larson69}Larson R B 1969 \mnras \textbf{145}

5415: 405--22

5416: 

5417: \bibitem[Larson and Tinsley(1978)]{Larson78}Larson R B and Tinsley

5418: B M 1978 \apj \textbf{219} 46--59

5419: 

5420: \bibitem[Larson \etal(1980)]{Larson80}Larson R B, Tinsley B M and

5421: Caldwell C N 1980 \apj \textbf{237} 692--707

5422: 

5423: \bibitem[Laurikainen \etal(2004)]{Laurikainen04}Laurikainen, E, Salo, H, and

5424: Buta, R 2004 \apj \textbf{607} 103--124

5425: 

5426: \bibitem[Lavery and Henry(1988)]{Lavery88}Lavery R J and Henry J P 1988

5427: \apj \textbf{330} 596--600

5428: 

5429: \bibitem[Lebreton(2001)]{Lebreton01}Lebreton  Y  2001 \araa \textbf{38}

5430: 35--77

5431: 

5432: \bibitem[Lee and Carney(1999)]{Lee99}Lee J-W and Carney B W 1999 \aj \textbf{118} 1373--89

5433: 

5434: \bibitem[Lee \etal(2007)]{Lee06}Lee H C, Worthey G, Trager S C and Faber S M 2007 \apj in press

5435: (\textit{Preprint} astro-ph/0605425)

5436: 

5437: \bibitem[Leitherer \etal(1999)]{Leitherer99}Leitherer C, Schaerer

5438: D, Goldader J D, Delgado R M, González R C, Kune D F, de Mello D F,

5439: Devost D and Heckman T M 1999 \apjs \textbf{123} 3--40

5440: 

5441: \bibitem[Levine \etal(2006)]{Levine06}Levine E S, Blitz L and Heiles,

5442: C 2006 \textit{Science} \textbf{312} 1773--7

5443: 

5444: \bibitem[Levy \etal(2007)]{Levy06}Levy  L, Rose J A, van Gorkom J

5445: and Chaboyer B 2007 \aj \textbf{133} 1104--24

5446: 

5447: \bibitem[Lo(2005)]{Lo05}Lo  K  Y  2005 \araa \textbf{43} 625--76

5448: 

5449: \bibitem[Lockman(2002)]{Lockman02}Lockman F J 2002 \textit{Seeing Through

5450: the Dust} (San Francisco: Astro. Soc. Pacific) \textit{ASP Conf.\ Proc.} vol

5451: 276 p 107

5452: 

5453: \bibitem[Lockman(2003)]{Lockman03}Lockman  F  J 2003 \apj \textbf{591}

5454: L33--36

5455: 

5456: \bibitem[Lombardi \etal(2005)]{Lombardi05}Lombardi  M  \etal\ 2005

5457: \apj \textbf{623} 42--56

5458: 

5459: \bibitem[Lucy(1974)]{Lucy74}Lucy  L R  1974 \aj \textbf{79} 745--54

5460: 

5461: \bibitem[Lynden-Bell(1967)]{Lynden67}Lynden-Bell D 1967 \mnras \textbf{136} 

5462: 101--21

5463: 

5464: \bibitem[Lynden-Bell and Wood(1968)]{Lynden68}Lynden-Bell D and Wood

5465: R 1968 \mnras \textbf{138} 495--525

5466: 

5467: \bibitem[Lynden-Bell and Kalnajs(1972)]{Lynden72}

5468: Lynden-Bell D and Kalnajs A J 1972 \mnras \textbf{157} 1--30

5469: 

5470: \bibitem[McConnachie \etal(2006)]{McConnachie06}McConnachie A W,

5471: Chapman S, Ibata R A, Ferguson A M N, Irwin M K, Lewis G F and Tanvir

5472: N R 2006 \apj \textbf{647} L25--8

5473: 

5474: \bibitem[McNamara \etal(2005)]{McNamara05}McNamara B R, Nulsen P

5475: E J, Wise M W, Rafferty R A, Carilli C, Sarazin C L and Blanton E

5476: L 2005 \nat \textbf{433} 45--7

5477: 

5478: \bibitem[MacArthur \etal(2003)]{MacArthur03}MacArthur L A, Courteau

5479: S and Holtzman J A 2003 \apj \textbf{582} 689--722

5480: 

5481: \bibitem[MacArthur, Courteau, Bell and Holtzman(2004)]{MacArthur04}MacArthur

5482: L, Courteau S, Bell E and Holtzman J A 2004 \apjs \textbf{152} 175--99 

5483: 

5484: \bibitem[Maccarone \etal(2004)]{Maccarone04}Maccarone  T  J, Fender

5485: R P and Tzioumis A K 2004 \textit{Preprint} astro-ph/0412014

5486: 

5487: \bibitem[Madgwick \etal(2003)]{Madgwick03}Madgwick

5488: D S, Somerville R, Lahav O, and Ellis R 2003 \mnras \textbf{343}

5489: 871--9

5490: 

5491: \bibitem[Majewski \etal(1996)]{Maj96}

5492: Majewski S R, Munn J A and Hawley S L 1996 \apj \textbf{459} L73--7

5493: 

5494: \bibitem[Majewski \etal(2003a)]{Majewski03}Majewski S  R, Skrutskie

5495: M F, Weinberg M D and Ostheimer J C 2003 \apj \textbf{599} 1082--1115

5496: 

5497: \bibitem[Majewski(2004)]{Majewski03a}Majewski S R  2004 \aj \textbf{128}

5498: 245--59

5499: 

5500: \bibitem[Malhotra(1995)]{Malhotra95}Malhotra S 1995 \apj \textbf{448}

5501: 138--48

5502: 

5503: \bibitem[Maraston \etal(2003)]{Maraston03}Maraston C, Greggio L,

5504: Renzini A, Ortolani S, Saglia R P, Puzia T H and Kissler-Patig M 2003

5505: \aap \textbf{400} 823--40

5506: 

5507: \bibitem[Marchant and Olson(1979)]{Marchant79}Marchant A B and Olson

5508: D W 1979 \apj \textbf{230} L157--9

5509: 

5510: \bibitem[Marcolini \etal(2005)]{Marcolini05}Marcolini A, Strickland

5511: D K, D'Ercole A, Heckman T M and Hoopes C G 2005 \mnras \textbf{362}

5512: 626--48

5513: 

5514: \bibitem[Marconi and Hunt(2003)]{Marconi03}Marconi A  and Hunt  L

5515:  K  2003 \apj \textbf{589} L21--4

5516: 

5517: \bibitem[Marinova and Jogee(2007)]{Marinova07}Marinova, I, and Jogee, S 2007

5518: \apj \textbf{659} 1176--1197

5519: 

5520: \bibitem[Markevitch \etal(2002)]{Markevitch02}Markevitch  M, Gonzalez,

5521: A H, Viklenin A, Murray S, Forman W, Jones C and Tucker W 2002 \apj

5522: \textbf{567} L27--31

5523: 

5524: \bibitem[Martins \etal(2005)]{Martins05}Martins L P, Gonzalez Delgado

5525: R M, Leitherer C, Cervino M and Hauschildt P 2005 \mnras \textbf{358}

5526: 49--65

5527: 

5528: \bibitem[Mateo(1998)]{Mateo98}Mateo  M  1998 \araa \textbf{36} 435--506

5529: 

5530: \bibitem[Mathewson \etal(1992)]{Mathewson92}Mathewson  D  S, Ford

5531: V L and Buchhorn M 1992 \apjs \textbf{81} 413--659

5532: 

5533: \bibitem[McGaugh \etal(2001)]{McGaugh01}McGaugh S, Rubin V C and

5534: de Blok W J G 2001 \aj \textbf{122} 2381--95

5535: 

5536: \bibitem[McGaugh(2004)]{McGaugh04}McGaugh S M 2004 \apj \textbf{609}

5537: 652--66

5538: 

5539: \bibitem[McGaugh(2005)]{McGaugh05a}McGaugh S S 2005 \apj \textbf{632}

5540: 859--71

5541: 

5542: \bibitem[McWilliam and Rich(1994)]{McWilliam94}McWilliam A and Rich

5543: R M 1994 \apjs \textbf{91} 749--91 

5544: 

5545: \bibitem[McWilliam(1997)]{McWilliam97}McWilliam A 1997 \araa \textbf{35}

5546: 503--56

5547: 

5548: \bibitem[Merritt(1993)]{Merritt93}Merritt  D  1993 \apj \textbf{413}

5549: 79--94

5550: 

5551: \bibitem[Merritt(2006)]{Merritt06}Merritt D 2006 \RPP \textbf{69} 2513--79

5552: 

5553: \bibitem[Mihos and Hernquist(1996)]{Mihos96}Mihos  J C and Hernquist

5554:  L  1996 \apj \textbf{464} 641--63

5555: 

5556: \bibitem[Mihos \etal(1999)]{Mihos99}Mihos  J  C, Spaans M and McGaugh

5557: S S 1999 \apj \textbf{515} 89--96

5558: 

5559: \bibitem[Minchin \etal(2005)]{Minchin05}Minchin  R  \etal\ 2005 \apj \textbf{622} L21--4

5560: 

5561: \bibitem[Michie(1963)]{Michie63}Michie R W 1963 \mnras \textbf{125} 127--39

5562: 

5563: \bibitem[Momany \etal(2004)]{Momany04}Momany  Y, Zaggia S R, Bonifacio

5564: P, Piotto G, De Angeli F, Bedin L R and Carraro G. 2004 \aap \textbf{421}

5565: L29--32

5566: 

5567: \bibitem[Monelli \etal(2003)]{Monelli03}Monelli M \etal 2003 \aj

5568: \textbf{126} 218--36

5569: 

5570: \bibitem[Moore \etal(1996)]{Moore96}Moore  B, Katz N, Lake G, Dressler

5571: A and Oemler A 1996 \nat \textbf{379} 613--6

5572: 

5573: \bibitem[Moorthey and Holtzman(2006)]{Moorthey05}Moorthy B K and

5574: Holtzman J A 2006 \mnras \textbf{371} 583--608

5575: 

5576: \bibitem[Morganti \etal(2006)]{Morganti06}Morganti R, de Zeeuw P

5577: T, Ooosterloo T A, McDermid R M, Krajnovi\'c D, Cappellari M, Kenn F

5578: and Weijmans A 2006 \mnras \textbf{371} 157--69

5579: 

5580: \bibitem[Moultaka \etal(2004)]{Moultaka04}Moultaka J, Ilovaisky S

5581: A, Prugniel P and Soubiran C 2004 \pasp \textbf{116} 693--8

5582: 

5583: \bibitem[Mouawad \etal(2005)]{Mouawad05}Mouawad N, Eckart A, Pfalzner

5584: S, Schödel R, Moultaka J and Spurzem R 2005 \textit{Astronomische Nachrichten}

5585: \textbf{326} 83--95

5586: 

5587: \bibitem[Mulchaey(2000)]{Mulchaey00}Muchaley J  2000 \araa \textbf{38}

5588: 289--335

5589: 

5590: \bibitem[Mulchaey and Regan(1997)]{Mulchaey97}Mulchaey, J S, and Regan, M W

5591: 1997 \apj \textbf{482} L135--L137

5592: 

5593: \bibitem[Napolitano \etal(2005)]{Napolitano05}Napolitano N  R, Capaccioli

5594: M, Romanowsky A J, Douglas N G, Merrifield M R, Kuijken K, Arnaboldi

5595: M, Gerhard O and Freeman K C 2005  \mnras \textbf{357} 691--706

5596: 

5597: \bibitem[Navarro and White(1994)]{Navarro94}Navarro J F and White

5598: S D M 1994 \mnras \textbf{267} 401--12

5599: 

5600: \bibitem[Navarro \etal(1997)]{Navarro97}Navarro  J  F,   Frenk  C

5601:  S  and White  S D  M  1997 \apj \textbf{490} 493--508

5602: 

5603: \bibitem[Newberg \etal(2002)]{Newberg02}Newberg H J \etal\  2002

5604: \apj \textbf{569} 245--74

5605: 

5606: \bibitem[Noordermeer(2006)]{Noordermeer06}Noordermeer  E  2006 \textit{The

5607: Distribution of Gas  Stars and Dark Matter in early-type disk galaxies}

5608: PhD thesis, University of Gronigen, The Netherlands

5609: 

5610: \bibitem[Noordermeer \etal(2007)]{Noordermeer07}

5611: Noordermeer E, van der Hulst J M, Sancisi R, Swaters R S and van 

5612: 	Albada T S 2007 \mnras \textit{376} 1515--46

5613: 

5614: \bibitem[Norris, Sharples and Kuntschner(2006)]{Norris06}Norris M

5615: A, Sharples R M and Kuntschner H 2006 \mnras \textbf{367} 815--24

5616: 

5617: \bibitem[Novati \etal(2006)]{Novati06}Novati S C, de Luca F, Jetzer P

5618: and Scarpetta G 2006 \aap \textbf{459} 407--14

5619: 

5620: \bibitem[O'Connell(1980)]{Oconnell80}O'Connell R W 1980 \apj \textbf{236}

5621: 430--40

5622: 

5623: \bibitem[O'Dea \etal(2004)]{Odea04}O'Dea C  P , Baum S A, Mack J

5624: and Koekemoer A M 2004 \apj \textbf{612} 131--51

5625: 

5626: \bibitem[Odenkirchen \etal(2003)]{Odenkirchen03}Odenkirchen M, Grebel

5627: E K, Dehnen W, Rix H-W, Yanny B, Newberg H J, Rockosi C M, Martínez-Delgado

5628: D, Brinkmann J and Pier J R 2003 \aj \textbf{126} 2385--407

5629: 

5630: \bibitem[Oey and Clarke(1997)]{Oey97}Oey M  S  and Clarke C J 1997

5631: \mnras \textbf{289} 570--88

5632: 

5633: \bibitem[Olling and Merrifield(1998)]{Ollig98}Olling  R  P and Merrifield

5634:  M  R  1998 \mnras \textbf{297} 943--52

5635: 

5636: \bibitem[Olling and Merrifield(2001)]{Ollig01}Olling  R  P  and Merrifield

5637:  M  R  2001 \mnras \textbf{326} 164--80

5638: 

5639: \bibitem[O'Neil \etal(2004)]{Oneil04}O'Neil  K, Bothun G, van Driel,

5640: W and Monnier Ragaigne D  \etal\ 2004 \aap \textbf{428} 823--35

5641: 

5642: \bibitem[Oort(1932)]{Oort32}Oort  J  1932 \textit{Bull. Astron. Inst.

5643: Netherlands} \textbf{6} 249--87

5644: 

5645: \bibitem[Oort(1961)]{Oort61}Oort J A 1961 \textit{The Distribution and Motion

5646: of Interstellar Matter in Galaxies} (New York: Benjamin), pp~3--12

5647: 

5648: \bibitem[Oppenheimer \etal(2001)]{Oppenheimer01}Oppenheimer  B  R,

5649: Hambly N C, Digby A P, Hodgkin S T and Saumon D 2001 \textit{Science}

5650: \textbf{292} 698--702

5651: 

5652: \bibitem[Palunas and Williams(2000)]{Palunas00}Palunas P  and Williams

5653:  T  B  2000 \aj \textbf{120} 2884--903

5654: 

5655: \bibitem[Panter \etal(2003)]{Panter03}Panter B, Heavens A F and Jimenez

5656: R 2003 \mnras \textbf{343} 1145--54

5657: 

5658: \bibitem[Pearce \etal(2001)]{Pearce01}Pearce F R, Jenkins A, Frenk

5659: C S, White S D M, Thomas P A, Couchman H M P, Peacock J A and Efstathiou

5660: G 2001 \mnras \textbf{326} 649--66

5661: 

5662: \bibitem[Peletier(1989)]{Peletier89}Peletier R F 1989, PhD Thesis

5663: University of Groningen, The Netherlands

5664: 

5665: \bibitem[Peletier \etal(1990)]{Peletier90}Peletier R F, Davies R L, Illingworth G D, 

5666: Davis L E and Cawson M 1990 \aj \textbf{100} 1091--142

5667: 

5668: \bibitem[Peletier \etal(1999)]{Peletier99}Peletier R F, Balcells

5669: M, Davies R L, Andrekakis Y, Vazdekis A, Burkert A and Prada F 1999

5670: \mnras \textbf{310} 703--16 

5671: 

5672: \bibitem[Persic \etal(1996)]{Persic96}Persic  M, Salucci P and Stel

5673:  F  1996 \mnras \textbf{281} 27--47

5674: 

5675: \bibitem[Peterson(2004)]{Peterson04}Peterson  B  M  2004 \apj \textbf{613}

5676: 682--99

5677: 

5678: \bibitem[Pfenninger and Revaz(2005)]{Pfenninger05}Pfenniger  D  and

5679: Revaz  Y  2005 \aap \textbf{431} 511--6

5680: 

5681: \bibitem[Pidopryhora \etal(2007)]{Pidopryhora06}Pidopryhora Y, Lockman

5682: F J and Shields J C 2007 \apj \textbf{656} 928--42

5683: 

5684: \bibitem[Pierce \etal(2006)]{Pierce06}Pierce M \etal\ 2006 \mnras

5685: \textbf{366} 1253--64

5686: 

5687: \bibitem[Pizagno \etal(2005)]{Pizagno05}Pizagno  J, Prada F, Weinberg

5688: D H, Rix H-W, Harbeck D, Grebel E K, Bell E F, Brinkmann J, Holtzman

5689: J and West A 2005 \apj \textbf{633} 844--56

5690: 

5691: \bibitem[Poggianti \etal(2004)]{Poggianti04}Poggianti  B, Bridges

5692: T J, Komiyama Y, Yagi M, Carter D, Mobasher B, Okamura S and Kashikawa

5693: N 2004 \apj \textbf{601} 197--213

5694: 

5695: \bibitem[Press and Schechter(1974)]{Press74}Press WH and Schechter

5696: P 1974 \apj \textbf{187} 425--38

5697: 

5698: \bibitem[Prochaska \etal(2007)]{Proc07}Prochaska L C, Rose J A,

5699: Caldwell N, Castilho B, Concannon K, Harding P, Morrison H and Schiavon

5700: R P 2007 \aj in press

5701: 

5702: \bibitem[Proctor and Samsom(2002)]{Proctor02}Proctor R N and Sansom

5703: A E 2002 \mnras \textbf{333} 517--43 

5704: 

5705: \bibitem[Quillen \etal(2006)]{Quillen06}Quillen A  C, Brookes M H,

5706: Keene J, Stern D, Lawrence C R and Werner M W 2006 \apj \textbf{645}

5707: 1092--1101

5708: 

5709: \bibitem[Rafferty \etal(2006)]{Rafferty06}Rafferty D A, McNamara

5710: B R, Nulsen P E J and Wise M W 2006 \apj \textbf{652} 216--31

5711: 

5712: \bibitem[Reddy \etal(2006)]{Reddy06}Reddy B E, Lambert D L. and Prieto

5713: C A 2006 \mnras \textbf{367} 1329--66

5714: 

5715: \bibitem[Reid \etal(2001)]{Reid01}Reid  I  N, Sahu K C and Hawley

5716: S L 2001 \apj \textbf{559} 942--7

5717: 

5718: \bibitem[Reid(2005)]{Reid05}Reid  I N 2005 \araa \textbf{43} 247--92

5719: 

5720: \bibitem[Renzini(1986)]{Renzini86}Renzini A 1986, \textit{Stellar Populations}

5721: (Cambridge: Cambridge University Press) p 213--23

5722: 

5723: \bibitem[Revnivtsev \etal(2004)]{Revnivtsev04}Revnivtsev  M  G, Churazov

5724: E M, Sazonov S Yu, Sunyaev R A, Lutovinov A A, Gilfanov M R, Vikhlinin

5725: A A, Shtykovsky P E and Pavlinsky M N 2004 \aap \textbf{425} L49--52

5726: 

5727: \bibitem[Rodgers and Paltoglou(1984)]{Rodg84}Rodgers A W and Paltoglou G 1984 \apj

5728: \textbf{283} L5--7

5729: 

5730: \bibitem[Rogstad \etal(1976)]{Rogstad76}Rogstad D H, Wright M and Lockhard I 1976 \apj

5731: \textbf{204} 703--11

5732: 

5733: \bibitem[Rogstad and Shostak(1978)]{Rogstad78}Rogstad D  H  and Shostak

5734:  G  S  1972 \apj \textbf{176} 315--21

5735: 

5736: \bibitem[Romanowsky \etal(2003)]{Romanowsky03}Romanowsky  A  J, Douglas

5737: N, Arnaboldi M, Kuijken K, Merrifield M R, Napolitano N R, Capaccioli

5738: M and Freeman K C 2003  \textit{Science} \textbf{301} 1696--8

5739: 

5740: \bibitem[Rose \etal(2005)]{Rose05}Rose J A, Arimoto N, Caldwell C

5741: N, Schiavon R P, Vazdekis A and Yamada Y 2005 \aj \textbf{129} 712--28

5742: 

5743: \bibitem[Rubin \etal(1978)]{Rubin78}Rubin  V  C,   Thonnard  N   and

5744: Ford  W  K  1978 \apj \textbf{225} L107--11

5745: 

5746: \bibitem[Rusin and Kochanek(2005a)]{Rusin05}Rusin  D  and Kochanek

5747: C  S  2005 \apj \textbf{623} 666--82

5748: 

5749: \bibitem[Rusin \etal(2005b)]{Rusin05a}Rusin  V  C,  Keeton C  R   and

5750: Winn  J  N  2005 \apj \textbf{627} L95--6

5751: 

5752: \bibitem[Ryden(2002)]{Ryden02}Ryden  B 2002 \textit{Introduction to

5753: Cosmology}  (New York: Addison-Wesley)

5754: 

5755: \bibitem[Sackett(1997)]{Sackett97}Sackett P  D  1997 \apj \textbf{483}

5756: 103--10

5757: 

5758: \bibitem[Salim \etal(2005)]{Salim05}Salim S \etal\  2005 \apj \textbf{619}

5759: L39--42 

5760: 

5761: \bibitem[Salpeter(1955)]{Salpeter55}Salpeter E E 1955 \apj \textbf{121}

5762: 161--7

5763: 

5764: \bibitem[Sanchez-Blazquez \etal(2006)]{Sanchez06}Sanchez-Blazquez

5765: P, Gorgas J, Cardiel N and Gonzalez J J 2006 \aap \textbf{457} 809--21

5766: 

5767: \bibitem[Sanders and McGaugh(2002)]{Sanders02}Sanders R H and McGaugh

5768: S S 2002 \araa \textbf{40} 263--317

5769: 

5770: \bibitem[Sargent \etal(1978)]{Sargent78}Sargent W L W, Young P J,

5771: Boksenberg A, Shortridge K, Lynds C R and Hartwick P D A 1978 \apj

5772: \textbf{221} 731--44

5773: 

5774: \bibitem[Saviance \etal(2000)]{Saviane00}Saviane I, Held E V and

5775: Bertelli G 2000 \aap \textbf{355} 56--68

5776: 

5777: \bibitem[Schaye(2004)]{Schaye04}Schaye  J  2004 \apj \textbf{609}

5778: 667--82

5779: 

5780: \bibitem[Schechter(1975)]{Schechter74}Schechter P L 1976 \apj

5781: \textbf{203} 297--306

5782: 

5783: \bibitem[Schechter and Gunn(1978)]{Schechter79}Schechter P L and

5784: Gunn J E 1979 \apj \textbf{229} 472--84

5785: 

5786: \bibitem[Schiavon \etal(2002a)]{Schiavon02a}Schiavon R P, Faber S

5787: M, Castilho B V and Rose J A 2002 \apj \textbf{580} 850--72

5788: 

5789: \bibitem[Schiavon \etal(2002b)]{Schiavon02b}Schiavon R P, Faber S

5790: M, Rose J A and Castilho, B V 2002 \apj \textbf{580} 873--86

5791: 

5792: \bibitem[Schiavon \etal(2004)]{Schiavon04}Schiavon R P, Caldwell

5793: C N and Rose J A 2004 \aj \textbf{127} 1513--30

5794: 

5795: \bibitem[Shlosman(2001)]{Shlosman01}Shlosman, I 2001 \textit{The Central

5796: Kiloparsec of Starbursts and AGN: The La Palma Connection} 

5797: (San Francisco: Astro. Soc. Pacific)

5798: \textit{ASP Conf.\ Proc.} vol 249) pp 55--77

5799: 

5800: \bibitem[Schmidt(1959)]{Schmidt65}Schmidt M 1959 \apj \textbf{129}

5801: 243--58

5802: 

5803: \bibitem[Schombert and Bothun(1987)]{Schombert87}Schombert J M and

5804: Bothun G D 1987 \aj \textbf{93} 60--73

5805: 

5806: \bibitem[Schr\"oder and Cuntz(2005)]{Schroder05}Schr\"oder K P and

5807: Cuntz M 2005 \apj \textbf{630} L73--6

5808: 

5809: \bibitem[Sch\"odel \etal(2003)]{Schodel03}Sch\"odel R, Ott T, Genzel R,

5810: Eckart A, Mouawad N and Alexander T 2003 \apj \textbf{596} 1015--34 

5811: 

5812: \bibitem[Schulz and Struck(2001)]{Schulz01}Schulz S and Struck C

5813: 2001 \mnras \textbf{328} 185--202

5814: 

5815: \bibitem[Schwarzschild(1958)]{Schwarzschild58}Schwarzschild M 1958

5816: \textit{Structure and Evolution of the Stars} (Princeton: Princeton

5817: Univ. Press)

5818: 

5819: \bibitem[Schwarzschild(1979)]{Schwarzschild79}Schwarzschild  M  1979

5820: \apj \textbf{232} 236--47

5821: 

5822: \bibitem[Searle \etal(1973)]{Searle73}Searle L, Sargent W L W and

5823: Bagnuolo W G 1973 \apj \textbf{179} 427--38

5824: 

5825: \bibitem[Searle and Zinn(1978)]{Searle78}Searle  L  and Zinn  H  1978

5826: \apj \textbf{225} 357--79

5827: 

5828: \bibitem[Seaton \etal(1994)]{Seaton94}Seaton M J, Yan Y, Mihalas

5829: D and Pradhan A K 1994 \mnras \textbf{266} 805--28

5830: 

5831: \bibitem[Seaton(1996)]{Seaton96}Seaton 1996 \mnras \textbf{279}

5832: 95--100

5833: 

5834: \bibitem[Sellwood and Wilkinson(1993)]{Sellwood93}Sellwood J A and Wilkinson A 1993 \RPP 

5835: \textbf{56} 173--256

5836: 

5837: \bibitem[S\'ersic(1968)]{Sersic68}S\'ersic  J L  1968 \textit{Atlas

5838: de Galaxias Australis} (Cordoba Argentina: Observatorio Astronomico)

5839: 

5840: \bibitem[Shetrone \etal(2003)]{Shetrone03}Shetrone M, Venn K A, Tolstoy

5841: E, Primas F, Hill V and Kaufer A 2003 \aj \textbf{125} 684--706 

5842: 

5843: \bibitem[Siegel \etal(2002)]{Siegel02}Siegel  M H, Majewski S R,

5844: Reid I N and Thompson I B 2002 \apj  \textbf{578} 151--75

5845: 

5846: \bibitem[Silk(1997)]{Silk77}Silk J 1997 \apj \textbf{481} 703--9

5847: 

5848: \bibitem[Solanes \etal(2001)]{Solanes01}Solanes J M, Manrique A, Garcia-Gomez C, 

5849: Gonzalez-Casado G., Giovanelli R and Haynes M P 2001 \apj \textbf{548} 97--113

5850: 

5851: \bibitem[Somerville and Kolatt(1999)]{Somerville99b}Somerville R

5852: S and Kolatt 1999 \mnras \textbf{305} 1--14

5853: 

5854: \bibitem[Somerville and Primack(1999)]{Somerville99}Somerville R

5855: S and Primack J R 1999 \mnras \textbf{310} 1087--1110

5856: 

5857: \bibitem[Spinrad and Taylor(1971)]{Spinrad71}Spinrad H and Taylor

5858: B J 1971 \apjs \textbf{22} 445--84

5859: 

5860: \bibitem[Springel and Hernquist(2003)]{Springel03}Springel  V  and

5861: Hernquist  L  2003 \mnras \textbf{339} 289--311

5862: 

5863: \bibitem[Springel \etal(2005)]{Springel05}Springel  V, Di Matteo

5864: T and Hernquist L 2005 \mnras \textbf{361} 776--794

5865: 

5866: \bibitem[Squires \etal(1996)]{Squires96}Squires G, Kaiser N, Babul A, Falhman G,

5867: Woods D, Neumann D and B\"ohringer H 1996 \apj \textbf{461} 572--86

5868: 

5869: \bibitem[Steinmetz and Navarro(1999)]{Steinmetz99}Steinmetz M and

5870: Navarro J F 1999 \apj \textbf{513} 555--60

5871: 

5872: \bibitem[Statler \etal(2004)]{Statler04}Statler T S, Emsellem E,

5873: Peletier R F and Bacon R 2004 \mnras \textbf{353} 1--14

5874: 

5875: \bibitem[Staniero(1997)]{Straniero97}Straniero O, Chieffi A and Limongi

5876: M1997 \apj \textbf{490} 425--36

5877: 

5878: \bibitem[Strickland \etal(2004)]{Strickland04}Strickland  D K, Heckman

5879: T M, Colbert E J M, Hoopes C G and Weaver K A 2004 \apjs \textbf{151}

5880: 193--236

5881: 

5882: \bibitem[Strutskie \etal(2006)]{Strutskie06}Strutskie M F \etal\ 

5883: 2006 \aj \textbf{131} 1163--83

5884: 

5885: \bibitem[Swaters, Sancisi and van der Hulst(1997)]{Swaters97}Swaters R A, Sancisi R and 

5886: van der Hulst J M 1997 \apj \textbf{366} 449--66

5887: 

5888: \bibitem[Thilker \etal(2005)]{Thilker05}Thilker D \etal\  2005 \apj

5889: \textbf{619} L79--92

5890: 

5891: \bibitem[Thomas \etal(2003)]{Thomas03}Thomas D,   Maraston S   and

5892: Bender  R  2003 \mnras \textbf{339} 897--911

5893: 

5894: \bibitem[Tonry and Schneider(1988)]{Tonry88}Tonry J and Schneider

5895: D P 1988 \aj \textbf{96} 807--15

5896: 

5897: \bibitem[Toomre(1981)]{Toomre81}Toomre A 1981 \textit{Structure and Evolution of Normal Galaxies} 

5898: (Cambridge: Cambridge University Press) pp~111--36

5899: 

5900: \bibitem[Toomre(1964)]{Toomre84}Toomre A 1964 \apj \textbf{139}

5901: 1217--38

5902: 

5903: \bibitem[Trager \etal(2000a)]{Trager00}Trager  S, Faber S M, Worthey

5904: G and Gonz\'alez J J 2000 \aj \textbf{119} 1645--76

5905: 

5906: \bibitem[Trager(2004)]{Trager04}Trager  S 2004 

5907: \textit{Origin and Evolution

5908: of the Elements} 

5909: (\textit{Carnegie Observatories Astrophys Series} vol 4) ed A McWilliam and M Rauch

5910: (Cambridge: Cambridge Univ Press) p~391

5911: 

5912: \bibitem[Tremaine \etal(2002)]{Tremaine02}Tremaine S  \etal\  2002

5913: \apj \textbf{574} 740--53

5914: 

5915: \bibitem[Tripicco and Bell(1995)]{Tripicco95}Tripicco M  and Bell

5916:  R  A  1995 \aj \textbf{110} 3035--49

5917: 

5918: \bibitem[Trujillo \etal(2001)]{Trujillo01}Trujillo L, Graham A W

5919: and Caon N 2001 \mnras \textbf{326} 869--76

5920: 

5921: \bibitem[Tsikoudi(1979)]{Tsikoudi79}Tsikoudi V 1979 \apj \textbf{234}

5922: 842--53

5923: 

5924: \bibitem[Tsujimoto and Shigeyama(1998)]{Tsujimoto98}Tsujimoto  T

5925:  and Shigeyama  T  1998 \apj \textbf{508} L151--4

5926: 

5927: \bibitem[di Tullio(1979)]{diTullio79}di Tullio G A 1979 \aap \textbf{37} 591--600

5928: 

5929: \bibitem[Tully and Fisher(1977)]{Tully76}Tully  R  B  and Fisher

5930:  J  R  1977 \aap \textbf{54} 661--73

5931: 

5932: \bibitem[Tully and Mohayaee(2004)]{Tully04}Tully  R  B and Mohayaee R 2004

5933: \textit{Outskirts of Galaxy Clusters: Intense Life in the Suburbs} ed A Diaferio 

5934: IAU Colloquium 195 p 205--11

5935: Torino Italy (Cambridge: Cambridge University Press)

5936: (\textit{Preprint} astro-ph/0404006)

5937: 

5938: \bibitem[Udalski \etal(1994)]{Udalski94}Udalski A, Szymanski M, Stanek K Z, Kaluzny J, Kubiak M,

5939: Mateo M, Krzeminski W, Paczynski B and Venkat R 1994 \textit{Acta Astron.} \textbf{44} 165--89

5940: 

5941: \bibitem[van Albada and Sancisi(1986)]{vanAlbada86}van Albada  T

5942:  S  and Sancisi  R  1986  \textit{Royal Soc. of London Philosophical

5943: Transactions Series A} \textbf{320} 447--64

5944: 

5945: \bibitem[van Dokkum \etal(1999)]{vanDokkum99}van Dokkum P G, Franx M,

5946: Fabricant D, Kelson D D and Illingworth G D 1999 \apj \textbf{520} L95--8

5947: 

5948: \bibitem[Vader \etal(1988)]{Vader88}Vader J P, Vigroux L, Lachieze-Rey M and Souviron J 1988 

5949: \aap \textbf{203}, 217--25

5950: 

5951: \bibitem[Valdes \etal(2004)]{Valdes04}Valdes F, Gupta R, Rose J A,

5952: Singh H P and Bell D J 2004 \apjs \textbf{152} 251--9

5953: 

5954: \bibitem[Valluri and Merritt(1998)]{Valluri98}Valluri M and Merritt

5955: D 1998 \apj \textbf{506} 686--711

5956: 

5957: \bibitem[Vazdekis \etal(1996)]{Vazdekis96}Vazdekis  A, Casuso E,

5958: Peletier R F and Beckman J E 1996 \apjs \textbf{106} 307--39

5959: 

5960: \bibitem[Veilleux \etal(2005)]{Veilleux05}Veilleux  S,   Cecil G

5961:   and Bland-Hawthorn  J  2005 \araa \textbf{43} 769--826

5962: 

5963: \bibitem[Venn \etal(2004)]{Venn04}Venn K A, Irwin, M, Shetrone M

5964: D, Tout C A, Hill V and Tolstoy E 2004 \aj\textbf{128} 1177--95 

5965: 

5966: \bibitem[van de Ven \etal(2003)]{vandeVen03}van de Ven G, Hunter

5967: C, Verolme E K and de Zeeuw P T 2003 \mnras \textbf{342} 1056--82

5968: 

5969: \bibitem[van Zee \etal(2004)]{vanZee04}van Zee L, Skillman E D and

5970: Haynes M P 2004 \aj \textbf{128} 121--36

5971: 

5972: \bibitem[Verheijen(2001)]{Verheijen01}Verheijen  M  A  W  2001 \apj

5973: \textbf{563} 694--715

5974: 

5975: \bibitem[Verolme \etal(2002)]{Verolme02}Verolme  E, Cappellari M,

5976: Copin Y, van der Marel R P, Bacon R, Bureau M, Davies R L, Miller

5977: B M and de Zeeuw P T 2002 \mnras \textbf{335} 517--25

5978: 

5979: \bibitem[Vollmer \etal(2006)]{Vollmer06}Vollmer B, Soida M, Otmianowska-Mazur

5980: K, Kenney J D P, van Gorkom J H and Beck R 2006 \aap \textbf{453}

5981: 883--93

5982: 

5983: \bibitem[Volonteri \etal(2003)]{Volonteri03}Volonteri  M, Madau P

5984: and Haardt F 2003 \apj \textbf{593} 661--6

5985: 

5986: \bibitem[Wallerstein \etal(1997)]{Wallerstein97}Wallerstein  G \etal\ 

5987: 1997 R\textit{ev  Mod  Phys}  \textbf{69} 995--1084

5988: 

5989: \bibitem[Wake \etal(2006)]{Wake06}Wake D A \etal\  2006 \mnras \textbf{372}

5990: 537--50

5991: 

5992: \bibitem[Weinberg and Blitz(2006)]{Weinberg06}Weinberg  M D  and

5993: Blitz  L  2006 \apj \textbf{641} L33--6

5994: 

5995: \bibitem[Weiner \etal(2001)]{Weiner01}Weiner B  J,   Sellwood  J

5996:  A   and Williams  T  B  2001 \apj \textbf{546} 931--51

5997: 

5998: \bibitem[Weiss \etal(2005)]{Weiss05}Weiss A, Serenelli A, Kitsikis

5999: A, Schlattl H and Christensen-Dalsgaard J 2005 \aap \textbf{441}

6000: 1129--33

6001: 

6002: \bibitem[White and Rees(1978)]{White78}White S D M and Rees M J 1978

6003: \mnras \textbf{183} 341--58

6004: 

6005: \bibitem[Whitmore and Schweizer(1995)]{Whitmore95}Whitmore  B  C

6006:  and Schweizer  F  1995 \aj \textbf{109} 960--80

6007: 

6008: \bibitem[Whyte \etal(2002)]{Whyte02}Whyte L F, Abraham R G, Merrifield M R, Eskridge P B, Frogel J A

6009: and Pogge R W 2002 \mnras \textbf{336} 1281--6

6010: 

6011: \bibitem[Willman \etal(2005)]{Willman05}Willman  B  \etal\  2005

6012: \apj \textbf{626} L85--8

6013: 

6014: \bibitem[Worthey(1994)]{Worthey94}Worthey  G  1994 \apjs \textbf{95}

6015: 107--49

6016: 

6017: \bibitem[Worthey(2004)]{Worthey04}Worthey, G 2004 \aj \textbf{128}

6018: 2826--37

6019: 

6020: \bibitem[Wu \etal(2005)]{Wu05}Wu H, Shao Z, Mo H J, Xia X and Deng Z 2005 \apj \textbf{622}

6021: 244--59

6022: 

6023: \bibitem[Wyithe and Loeb(2003)]{Wyithe03}Wyithe  J  S  B  and Loeb

6024:  A  2003 \apj \textbf{595} 614--23

6025: 

6026: \bibitem[Yoachim and Dalcanton(2006)]{Yoachim06}Yoachim P and Dalcanton

6027: J J 2006 \aj 131 226--49

6028: 

6029: \bibitem[Yoon, Yi and Lee(2006)]{Yoon06}Yoon S-J, Yi S K and Lee

6030: Y-W 2006 Science \textbf{311} 1129--32

6031: 

6032: \bibitem[York \etal(2000)]{York00}York D G \etal~2000 \aj \textbf{120}

6033: 1579--87

6034: 

6035: \bibitem[de Zeeuw \etal(2002)]{deZeeuw02}de Zeeuw P T  \etal\  2002

6036: \mnras \textbf{329} 513--30

6037: 

6038: \bibitem[Zhang \etal(2001)]{Zhang01}Zhang  J, Fall S M and Whitmore

6039: B C 2001 \apj \textbf{561} 727--50

6040: 

6041: \bibitem[Zoccali \etal(2003)]{Zoccali03}Zoccali M, Renzini A, Ortolani

6042: S, Greggio L, Saviane I, Cassisi S, Rejkuba M, Barbuy B, Rich R M

6043: and Bica E 2003 \aap \textbf{399} 931--56

6044: 

6045: \bibitem[Zwicky(1937)]{Zwicky37}Zwicky  F 1937 \apj \textbf{86}

6046: 217--46

6047: \end{harvard}

6048: 

6049: \end{document}

6050: